Thesis Defense


  Ph.D. Final Defense

Title: Privacy Preservation for Cloud-Based Data Sharing and Data Analytics

Speaker: Yao Zheng

Date: Nov. 14, 2016, 10am-12pm, Room NVC 207


  Ph.D. Preliminary Exam Defense

Title: Trust-based Service Management of Internet of Things Systems and Its Applications

Speaker: Jia Guo

Nov. 9, 2016, 9-11am, Room NVC 103


  Ph.D. Research Defense

Title: Privacy Preservation for Cloud-Based Data Sharing and Data Analytics

Speaker: Yao Zheng

Date: Sept. 2, 2016, 11am-1pm, Room NVC 320


  Ph.D. Final Defense

Title: Attack and Defense with Hardware-Aided Security

Speaker: Ning Zhang

Date: June 21, 2016, 9:00-11:00am, Room NVC 320


  Ph.D. Final Defense

Title: Spatio-temporal Event Detection and Forecasting in Social Media

Speaker: Liang Zhao

Date: July 22, 2016, 10:00am, T3


  MS Thesis Defense

Title: Looks Good to Me: Authentication for Augmented Reality

Speaker: Ethan D. Gaebel

Date: May 1, 2016, 2:00-3:00am, Room NVC 320


  Ph.D. Final Defense

Title: Search over Encrypted Data in Cloud Computing

Speaker: Bing Wang

Date: Apr. 28, 2016, 1:00-3:00pm, Room NVC T3


  Ph.D. Research Defense

Title: Spatio-temporal Event Detection and Forecasting in Social Media

Speaker: Liang Zhao

Date: Apr. 27, 2016, 9:00-11:00am, Room NVC 320


  Ph.D. Final Defense

Title: Trust-Based Service Management for Service-Oriented MANETs and Its Application to Service Composition and Task Assignment with Multi-Objective Goals

Speaker: Yating Wang

Date: Apr. 25, 2016, 9:30-11:30am, Room NVC T3


  Ph.D. Research Defense

Title: Trust-Based Service Management for Service-Oriented MANETs and Its Application to Service Composition and Task Assignment with Multi-Objective Goals

Speaker: Yating Wang

Date: Feb. 12, 2016, 9:30-11:30am, Room NVC 320


  Ph.D. Preliminary Exam Defense

Title: Privacy Preservation for Cloud-Based Data

Speaker: Yao Zheng

Feb. 5, 2016, 2-4pm, Room NVC 320


  Ph.D. Research Defense

Title: Protecting Data Privacy in Cloud Computing Applications

Speaker: Bing Wang

Date: Jan. 25, 2016, 2:30-4:30pm, Room NVC 320


  Ph.D. Research Defense

Title: Attack and Defense with Hardware Aided Security

Speaker: Ning Zhang

Date: Jan. 22, 2016, 2-4pm, Room NVC 320


  Ph.D. Preliminary Exam Defense

Title: Spatio-temporal Event Detection and Forecasting in Social Media

Speaker: Liang Zhao

Date: Oct. 22, 2015, 10:00-12:00am, Room NVC 320


  Ph.D. Preliminary Exam Defense

Title: Protecting Data Privacy in Cloud Computing

Speaker: Bing Wang

Date: Sept. 9, 2015, 9:30-11:30am, Room NVC 207


  Ph.D. Preliminary Exam Defense

Title: Attack and Defense with Hardware Aided Security

Speaker: Ning Zhang

Date: March 27, 2015, 9-11am, Room NVC 203


  Ph.D. Preliminary Exam Defense

Title: Trust Management for Service-Oriented MANETs and Its Application to Service Composition and Task Assignment with Multi-Objective Goals

Speaker: Yating Wang

Date: Sept 5, 2014, 1-4pm, Room NVC 320


  Ph.D. Final Defense

Title: Toward Security Enhanced Cognitive Radio Networks

Speaker: Qiben Yan

Date: June 23, 2014, 10-12am, Room NVC 207


  Ph.D. Final Defense

Title: Data Analysis in Spatial Contexts

Speaker: Raimundo F. Dos Santos Jr.

Date: May 19, 2014, 10am, Room NVC 320


  Ph.D. Final Defense

Title: Dynamic Redundancy Management of Multisource Multipath Routing Integrated with Voting-based Intrusion Detection in Wireless Sensor Networks

Speaker: Hamid Al-Hamadi

Date: April 2, 2014, 2pm-4pm, Room NVC 320


  Ph.D. Research Defense

Title: Data Analysis in Spatial Contexts

Speaker: Raimundo F. Dos Santos Jr.

Date: Feb. 21, 2014, 10am, Room NVC 320


  Ph.D. Research Defense

Title: Toward Security Enhanced Cognitive Radio Networks

Speaker: Qiben Yan

Date: Jan. 23, 2014, 10-12am, Room NVC 351


  Ph.D. Research Defense

Title: Dynamic Redundancy Management of Multisource Multipath Routing Integrated with Voting-based Intrusion Detection in Wireless Sensor Networks

Speaker: Hamid Al-Hamadi

Date: Nov. 19, 2013, 10am-12pm, Room NVC 320


  Ph.D. Final Defense

Title: Factors Affecting the Design and Use of Reusable Components

Speaker: Reghu Anguswamy

Date: Date: June 13, 2013, 10:30am-12:30pm, Room NVC T3


  Ph.D. Preliminary Exam Defense

Title: Toward Security Enhanced Cognitive Radio Networks

Speaker: Qiben Yan

Date: May 28, 2013, 10-12am, Room NVC 351


  Ph.D. Final Defense

Title: Prediction and Anomaly Detection Techniques for Spatial Datasets

Speaker: Xutong Liu

Date: May 7, 2013, 11am-1pm, Room NVC 351


  Ph.D. Final Defense

Title: Dynamic Trust Management for Mobile Networks

Speaker: Fenye Bao

Date: May 6, 2013, 5-7pm, Room NVC 351


  MS Thesis Defense

Title: A Military Planning Methodology for Conducting Cyber Attacks on Power Grid

Speaker: Mehmet Saglam

Date: May 5, 2013, 11:00am-12:30pm, Room NVC 207


  MS Thesis Defense

Title: Robust Prediction of Large Spatio-Temporal Datasets

Speaker: Yang Chen

Date: May 1, 2013, 10:30-12:00am, Room NVC 313


  MS Thesis Defense

Title: Dual Path PKI for Secure Aircraft Data Communication

Speaker: Alexander K. Buchholz

Date: April 29, 2013, 2-4pm, Room NVC 351


  Ph.D. Final Defense

Title: Design and Analysis of Intrusion Detection Protocols in Cyber Physical Systems

Speaker: Robert Mitchell

Date: April 5, 2013, 10am, Room NVC 401


 Ph.D Research Defense

Title: Prediction and Anomaly Detection Techniques for Spatial Datasets

Speaker: Xutong Liu

Date: Jan. 24, 2013, 10-12am, Room NVC T3


  MS Thesis Defense

Title: Automated Seed Point Selection in Confocal Image Stacks of Neuron Cells

Speaker: Gregory P. Bilodeau

Date: Jan. 2, 2013, 2-3pm, Room NVC 325


  Ph.D. Research Defense

Title: A Study of Factors Affecting the Design and Use of Reusable Components

Speaker: Reghu Anguswamy

Date: Dec. 14, 2012, 11am, Room NVC 221


  Ph.D. Preliminary Exam Defense

Title: Dynamic Redundancy Management of Multisource Multipath Routing Integrated with Voting-based Intrusion Detection in Wireless Sensor Networks

Speaker: Hamid Al-Hamadi

Date: Dec. 7, 2012, 2pm, Room NVC 401


  Ph.D. Final Defense

Title: Efficient Algorithms for Mining Large Spatio-Temporal Data

Speaker: Feng Chen

Date: Nov. 30, 2012, 3pm, Room NVC 207


  Ph.D. Research Defense

Title: Dynamic Trust Management for Mobile Networks and Its Applications

Speaker: Fenye Bao

Date: Nov. 28, 2012, 4pm, Room NVC T3


  Ph.D. Research Defense

Title: Design and Analysis of Intrusion Detection Protocols in Cyber Physical Systems

Speaker: Robert Mitchell

Date: Nov. 15, 2012, 2pm, Room NVC 401


  Ph.D. Preliminary Exam Defense

Title: Data Analysis in Spatial Contexts

Speaker: Raimundo F. Dos Santos Jr.

Date: Sept. 7, 2012, 10am, Room NVC 351


  Ph.D. Preliminary Exam Defense

Title: A Study of Factors Affecting the Design and Use of Reusable Components

Speaker: Reghu Anguswamy

Date: May 9, 2012, 2pm, Room NVC 204


  Ph.D. Research Defense

Title: Efficient Algorithms for Mining Large Spatio-Temporal Data

Speaker: Feng Chen

Date: May 10, 2012, 2pm, Room NVC 351


  Ph.D. Final Defense

Title: Integrated Mobility and Service Management for Network Cost Minimization in Wireless Mesh Networks

Speaker: Yinan Li

Date: April. 30, 2012, 11am, Room NVC 351


  Ph.D. Final Defense

Title: Formal Specification and Verification of Data-Centric Web Services

Speaker: Iman Saleh Moustafa

Date: Feb. 10, 2012, 2:30pm, Room NVC 204


  Ph.D. Preliminary Exam Defense

Title: Dynamic Trust Management for Mobile Networks and Its Applications

Speaker: Fenye Bao

Date: Nov. 21, 2011, 3pm, Room NVC 325


  Ph.D. Preliminary Exam Defense

Title: Outlier Detection Techniques in Spatial Dataset

Speaker: Xutong Liu

Date: Nov. 16, 2011, 2pm, Room NVC 351


  Ph.D. Preliminary Exam Defense

Title: Design and Analysis of Intrusion Detection Protocols in Cyber Physical Systems

Speaker: Robert Mitchell

Date: Nov. 10, 2011, 10am, Room NVC 351


  Ph.D. Research Defense

Title: Formal Specification and Verification of Data-Centric Web Services

Speaker: Iman Saleh Moustafa

Date: Oct. 24, 2011, 11am, Room NVC 204


  MS Thesis Defense

Title: Preserving Unique References in Java Lists

Speaker: Daniel Wayne Smith

Date: Dec. 7, 2010


  Ph.D. Preliminary Exam Defense

Title: Integrated Mobility and Service Management for Cost Minimization in Wireless Mesh Networks

Speaker: Yinan Li

Date: October 27, 2010, 10am, Room 320


  Ph.D. Preliminary Exam Defense

Title: On Local Based Algorithms for Mining Spatial Data

Speaker: Feng Chen

Date: October 5, 2010


 MS Thesis Defense

Title: Mining Social Tags to Predict Mashup Patterns

Speaker: Khaled Goarany

Date: Sept. 10 2010


  Ph.D. Final Defense

Title: Efficient Algorithms for Mining Data Streams

Speaker: Arnold P. Boedihardjo

Date: August 10, 2010, 2pm, Room 351


  MS Thesis Defense

Title: Comparison of Domain Vocabularies Across Domain Experts

Speaker: Chaitanya Nemallapudi

Date: August 2, 2010, 2pm, Room 325


  MS Thesis Defense

Title: A Sufficient Set of Mutation Operators for Structured Query Language (SQL)

Speaker: Donald W. McCormick II

Date: Apr. 28, 2010, 2pm, Room 325


  Ph.D. Research Defense

Title: Efficient Algorithms for Mining Data Streams

Speaker: Arnold P. Boedihardjo

Date: Jan. 27, 2010, 2pm, Room 313


  PhD Preliminary Exam Defense

Title: Specification and Verification of Data-Centric Web Services

Speaker: Iman Saleh Moustafa

Date: Dec. 4, 2009, 10am, Room 207


 Ph.D. Final Defense

Title: Efficient Concurrent Operations in Spatial Databases

Speaker: Jing Dai

Date: Sept. 4, 10am, 2009, Room 351


 Ph.D. Final Defense

Title: Mobility and Service Management for Future All-IP based Wireless Networks

Speaker: Weiping He

Date: March 20, 10am, 2009, Room 322


  PhD Preliminary Exam Defense

Title: Efficient Algorithms for Mining Data Streams

Speaker: Arnold P. Boedihardjo

Date: Jan. 28, 10am, 2009. Room 313


 Master's Thesis Defense

Title: A Comparison of Statistical Filtering Methods for Automatic Term Extraction for the Purpose of Domain Engineering

Speaker: Jason Tilley

Date: Dec. 22, 1pm, 2008. Room 325


 Master's Thesis Defense

Title: Tree Component alternatives to the Composite Design Pattern

Speaker: Arun Sudhir

Date: Dec. 3, 11pm, 2008. Room 324


 Ph.D. Research Defense

Title: MOBILITY AND SERVICE MANAGEMENT FOR FUTURE ALL-IP BASED WIRELESS NETWORKS

Speaker: Weiping He

Date: Dec. 3, 2pm, 2008. Room 325


 Ph.D. Final Defense

Title: A Class of Call Admission Control Algorithms for Resource Management and Reward Optimization for Servicing Multiple QoS Classes in Wireless Networks and Its Applications

Speaker: Okan Yilmaz

Date: Nov. 17th 2008, 2pm, NVC Room 325


 Ph.D. Final Defense

Title: Design and Analysis of QoS-Aware Key Management and Intrusion Detection Protocols for Secure Mobile Group Communications in Wireless Networks

Speaker: Jin-Hee Cho

Date: Nov. 12th 2008, 1pm, NVC Room 313


 Ph.D. Research Defense

Title: A Class of Call Admission Control Algorithms for Resource Management and Reward Optimization for Servicing Multiple QoS Classes in Wireless Networks and Its Applications

Speaker: Okan Yilmaz

Date: August 14th 2008, 2pm, NVC Room 324


 Ph.D. Research Defense

Title: Design and Analysis of QoS-Aware Key Management and Intrusion Detection Protocols for Secure Mobile Group Communications in Wireless Networks

Speaker: Jin-Hee Cho

Date: May 29th 2008, 2pm, NVC Room 320


 Master's Thesis Defense

Title: Incorporating Design Knowledge into Genetic Algorithm-based White-Box Software Test Case Generators

Speaker: Matthew C. Makai

Date: April 24th 2008, 2pm, NVC Room 322


 PhD Final Defense

Title: Design and Analysis of Adaptive Fault Tolerant QoS Control Algorithms for Query Processing in Wireless Sensor Networks

Speaker: Anh Phan Speer

Date: April 17, 2008, 1pm, NVC Room 320


 PhD Research Defense

Title: Design and Analysis of Adaptive Fault Tolerant QoS Control Algorithms for Query Processing in Wireless Sensor Networks

Speaker: Anh Phan Speer

Date: Dec. 20 2007, 11am, NVC Room 320


 PhD Final Defense

Title: Service-Oriented Sensor-Actuator Networks

Speaker: Abdelmounaam Rezgui

Date: November 29th 2007, 10am, NVC Room 106


 PhD Final Defense

Title: Multi-channel Mobile Access to Web Services

Speaker: Xu Yang

Date: November 26th 2007, 5pm, NVC Room 111


 MS Defense

Title: Empirical Analysis of Value and Reference Semantics

Speaker: Neha Khedekar

Date: August 10th 2007, 2pm, NVC Room 314


 MS Defense

Title: A component-based approach to proving the correctness of the Schorr-Waite algorithm

Speaker: Amrinder Singh

Date: August 9th 2007, 2pm, NVC Room 325


 PhD Research Defense

Title: Support for Subjective Views in Collaborative Virtual Environments

Speaker: Jianghui Ying

Date: July 19 2007,5pm, NVC Room 111


 PhD Research Defense

Title: Multi-channel Mobile Access to Web Services

Speaker: Xu Yang

Date: June 14, 5pm, 2007, Room 103


 PhD Preliminary Exam Defense

Title: Design and Analysis of QoS-Aware Key Management and Intrusion Detection Protocols for Secure Mobile Group Communications in Wireless Networks

Speaker: Jin-Hee Cho

Date: May 9, 10am, 2007. Room 320


 PhD Preliminary Exam Defense

Title: A Framework for Resource and Pricing Management for Revenue Optimization with QoS Guarantees for Multiple Service Classes in Wireless Networks

Speaker: Okan Yilmaz

Date: March 21, 2pm, 2007. Room 207


 PhD Preliminary Exam Defense

Title: MOBILITY AND SERVICE MANAGEMENT FOR FUTURE ALL-IP BASED WIRELESS NETWORKS

Speaker: Weiping He

Date: Dec. 12, 2pm, 2006. Room 324


 PhD Final Exam Defense

Title: Abnormal Pattern Recognition in Spatial Data

Speaker: Yufeng Kou

Date: Nov. 29, 10am, 2006. Room 103


 PhD Preliminary Exam Defense

Title: Efficient Concurrent Operations in Spatial Databases

Speaker: Jing Dai

Date: Oct. 30, 10am, 2006. Room 103


 PhD Preliminary Exam Defense

Title: Design and Analysis of Adaptive Fault Tolerant QoS Control Algorithms for Query Processing in Wireless Sensor Networks

Speaker: Anh Phan Speer

Date: Oct. 20, 1pm, 2006. Room 320


 Master's Thesis Defense

Title: Software Agents for DLNET Content Review - Study and Experiment

Speaker: Seema Mitra

Date: August 31, 1pm, 2006. Room 204 (max cap. 30)


 Master's Thesis Defense

Title: The Design and Implementation of the Tako Language and Compiler

Speaker: Jyotindra Vasudeo

Date: May 5th, 12:30pm, 2006


 Master's Thesis Defense

Title: Managing Changes to Service Oriented Enterprises

Speaker: M. Salman Akram


 PhD Thesis Defense

Title: Design and Analysis of Algorithms for Efficient Location and Service Management in Mobile Wireless Systems

Speaker: Baoshan Gu

Date: Sep. 30th 1:00pm-3:00pm
 
 
 
 





You are cordially invited to attend Yao Zheng's Ph.D. Final Defense
Nov. 14, 2016, 10am-12pm, Room NVC 207



Speaker: Yao Zheng
Advisor: Prof. Wenjing Lou


Title: Privacy Preservation for Cloud-Based Data Sharing and Data Analytics

Abstract:


Data privacy is a globally recognized human right for individuals to protect their sensitive personal information stored on computer systems. As communication technology progresses, the means to protect data privacy must also evolve to address new challenges come into view. Our research goal in this thesis is to develop privacy models and privacy-enhancing technologies for emerging cloud-based data services, in particular privacy-preserving algorithms and protocols for cloud-based data sharing and data analytics.

The cloud computing architecture has enabled users to store, process, and communicate their personal information through third-party cloud services. It has also raised privacy issues regarding losing control over data, mass harvesting of information, and disclosure of personal content. Above all, the main concern is the absence of clear privacy de finitions for cloud-based data services. Currently, the cloud service providers either abide by the principle of third-party doctrine, and off ers little privacy protection to users' personal data, or adopt the transparency-and-choice approach, and create complicate privacy statements, which are ignored by users.

In this regard, our research has three main contributions. First, to study users' privacy expectations, we conceptually divide personal data into two categories, i.e., active data and passive data. The active data refer to data knowingly generated by users and shared through the cloud (examples include personal health records, location check-in, etc.). The passive data refer to data routinely generated about users and automatically collected by machines (examples include inferences about users' purchasing habits and social network relationships, faceprint templates derived from user's photos, etc.).

Second, we propose two succinct and clear privacy definitions, namely individual control and use limitation, for the two data categories. The individual control model emphasizes users' capability to govern the access of their (active) data stored in the cloud. The use limitation model emphasizes users' expectation to remain anonymous when their (passive) data are aggregated and analyzed in the cloud.

Finally, we investigate various techniques to achieve these goals, in the context of four cloud-based data services: personal health record sharing, location-based proximity test, recommender systems with collaborative filtering, and face tagging with deep learning. For the first case, we develop a cryptography based approach to enforce fi ne-grained access control to users' health record. For the second case, we develop an obfuscation based approach to achieve location-specific user selection. For latter two cases, we develop distributed learning algorithms and interactive data release mechanisms to prevent large scale data harvesting.

We further combine them with differential privacy techniques, such as additive noise and sparse vector, to achieve user anonymity. The picture that is emerging from the above work is a bleak one. Regarding to personal data, the reality is we can no longer control them all. As communication technology evolve, the scope of personal data has expanded beyond local, discrete silos, and integrated into the Internet. The traditional privacy model must be updated to capture the society's new privacy expectation. Because privacy is a particularly nuanced problem that is subject to social norms, there is also no one size t all solution. While some cases can be salvaged either by cryptography or by other means, in others a rethinking of the trade-o s between utility and privacy appears to be necessary.


You are cordially invited to attend Jia Guo's Ph.D. Preliminary Exam Defense
Nov. 9, 2016, 9-11am, Room NVC 103



Speaker: Jia Guo
Advisor: Prof. Ing-Ray Chen


Title: Trust-based Service Management of Internet of Things Systems and Its Applications

Abstract:


The goals of the dissertation research are to develop trust protocols utilizing general trust management techniques for distributed, centralized, and hybrid IoT applications, verify desirable properties including solution quality, accuracy, convergence, resiliency, and scalability having been achieved, prove the designed protocols outperform contemporary trust protocols, if any exists, and prove the validity of our trust protocols with real-world IoT applications running in distributed, centralized, and hybrid IoT environments.


You are cordially invited to attend Yao Zheng's Ph.D. Research Defense
Sept. 2, 2016, 11am-1pm, Room NVC 320



Speaker: Yao Zheng
Advisor: Prof. Wenjing Lou


Title: Privacy Preservation for Cloud-Based Data Sharing and Data Analytics

Abstract:


Data privacy is a globally recognized human right for individuals to protect their sensitive personal information stored on computer systems. As communication technology progresses, the means to protect data privacy must also evolve to address new challenges come into view. Our research goal in this thesis is to develop privacy models and privacy-enhancing technologies for emerging cloud-based data services, in particular privacy-preserving algorithms and protocols for cloud-based data sharing and data analytics.

The cloud computing architecture has enabled users to store, process, and communicate their personal information through third-party cloud services. It has also raised privacy issues regarding losing control over data, mass harvesting of information, and disclosure of personal content. Above all, the main concern is the absence of clear privacy de finitions for cloud-based data services. Currently, the cloud service providers either abide by the principle of third-party doctrine, and off ers little privacy protection to users' personal data, or adopt the transparency-and-choice approach, and create complicate privacy statements, which are ignored by users.

In this regard, our research has three main contributions. First, to study users' privacy expectations, we conceptually divide personal data into two categories, i.e., active data and passive data. The active data refer to data knowingly generated by users and shared through the cloud (examples include personal health records, location check-in, etc.). The passive data refer to data routinely generated about users and automatically collected by machines (examples include inferences about users' purchasing habits and social network relationships, faceprint templates derived from user's photos, etc.).

Second, we propose two succinct and clear privacy definitions, namely individual control and use limitation, for the two data categories. The individual control model emphasizes users' capability to govern the access of their (active) data stored in the cloud. The use limitation model emphasizes users' expectation to remain anonymous when their (passive) data are aggregated and analyzed in the cloud.

Finally, we investigate various techniques to achieve these goals, in the context of four cloud-based data services: personal health record sharing, location-based proximity test, recommender systems with collaborative filtering, and face tagging with deep learning. For the first case, we develop a cryptography based approach to enforce fi ne-grained access control to users' health record. For the second case, we develop an obfuscation based approach to achieve location-specific user selection. For latter two cases, we develop distributed learning algorithms and interactive data release mechanisms to prevent large scale data harvesting.

We further combine them with differential privacy techniques, such as additive noise and sparse vector, to achieve user anonymity. The picture that is emerging from the above work is a bleak one. Regarding to personal data, the reality is we can no longer control them all. As communication technology evolve, the scope of personal data has expanded beyond local, discrete silos, and integrated into the Internet. The traditional privacy model must be updated to capture the society's new privacy expectation. Because privacy is a particularly nuanced problem that is subject to social norms, there is also no one size t all solution. While some cases can be salvaged either by cryptography or by other means, in others a rethinking of the trade-o s between utility and privacy appears to be necessary.


You are cordially invited to attend Ning Zhang's Ph.D. Final Defense
July 22, 2016, 10am, T3



Speaker: Ning Zhang
Advisor: Prof. Wenjing Lou


Title: Attack and Defense with Hardware-Aided Security

Abstract:


Riding on recent advances in computing and networking, our society is now experiencing the evolution into the age of information. While the development of these technologies brings great value to our daily life, the lucrative reward from cyber-crimes has also attracted criminals. Unwanted software is delivered to the victims through various methods such as remote exploitation and social engineering. As computing continues to play an increasing role in the society, security has become a pressing issue. Failures in computing systems could result in loss of infrastructure or human life, as demonstrated in both academic research and production environment. With the continuing widespread of malicious software and new vulnerabilities revealing every day, protecting the heterogeneous computing systems across the Internet has become a daunting task.

My approach to this challenge consists of two directions. The first direction aims to gain better understanding of the inner working of both attacks and defenses in the cyber environment. It was said in Art of War by Sun Tzu, ”If you know your enemies and know yourself, you will not be imperiled in a hundred battles”. I share the same belief that in order to design systems that are resistant to cyber attackers, it is necessary to understand how cyber-attack works. Under this direction, part of my research focus on examining the evolution of cyber-attack. Meanwhile, my other direction is designing secure execution environments using security features offered by hardware. Hardware-aided security offers the opportunity to bootstrap a secure environment even when the software environment is compromised. Under this direction, my research focuses on providing trusted environment in an adversarial setting with the presence of powerful attackers.

The next generation of computing consists of powerful elastic cloud and mobile devices inter-connecting via fast and reliable networks. Therefore to provide a complete solution, our work spans both the cloud and mobile endpoint.


You are cordially invited to attend Liang Zhao's Ph.D. Final Defense
June 21, 2016, 9:00-11:00am, Room NVC 320



Speaker: Liang Zhao
Advisor: Prof. Chang-Tien Lu


Title: Spatio-temporal Event Detection and Forecasting in Social Media

Abstract:


Nowadays, knowledge discovery on social media has attracted growing interests. Social media, far more than a communication tool, functions as social sensors for our society. Hundreds of millions of users collectively post millions of tweets every hour, discussing a variety of content ranging from everyday feelings to erent from traditional documents, social media exhibits many other interesting properties: 1) Timeliness of messages: Unlike traditional media that take hours or days to publish, tweets can be posted instantly utilizing portable mobile devices; 2) Ubiquity of social sensors: Tweets reflect the public mood and trends, which could be the determinants of future social events; and 3) Availability of geoinformation: Twitter users provide rich location information in proles, texts, and geo-tags.

This thesis focuses on the development of methods for social media-based spatiotemporal event topics and assumptions. Four methods are proposed, including dynamic query expansion for event detection, a generative framework for event forecasting, multi-task learning for spatiotemporal event forecasting, and deep learning based epidemics modeling for forecasting. For the first method, existing solutions for spatiotemporal event detection are mostly supervised and lack flexibility to handle dynamic keywords of social media. The proposed contributions of our work include: (1) Develop an unsupervised framework; (2) Design a novel dynamic query expansion (DQE) method; and (3) Propose an innovative local modularity spatial scan (LMSS) algorithm.

For the second method, traditional solutions for spatiotemporal event forecasting were mostly designed based on simple assumptions and cannot handle the complexity of our task, such as capture the spatiotemporal context, model mixed-type observations, and utilize prior geographical knowledge. The proposed contributions of our work for this task include: (1) Propose a novel generative model for spatial event algorithm for model parameter inference; and (3) Develop a new sequence likelihood calculation method. For the third method, traditional solutions cannot jointly consider the heterogeneity and effectively handle the dynamics of social media data. The proposed contributions for this work include: (1) Formulate a multi-task learning framework for event forecasting; and (2) Jointly model static and dynamic terms. (3) Develop efficient parameter optimization algorithms. For the last method, existing work on epidemics modeling either cannot ensure the timeliness of disease surveillance, or cannot effectively characterize the underlying epidemics mechanism. The contributions of this work include: Propose a novel integrated framework for computational epidemiology and social media mining (2) Develop a semi-supervised multilayer perceptron (MLP) for mining epidemic features, and (3) Design an online training algorithm.


You are cordially invited to attend Ethan D. Gaebel's MS Thesis Defense
May 1, 2016, 2:00-3:00am, Room NVC 320



Speaker: Ethan D. Gaebel
Advisor: Prof. Wenjing Lou


Title: Looks Good to Me: Authentication for Augmented Reality

Abstract:


Augmented reality is poised to become the next dominant computing paradigm over the course of the next decade. With the three-dimensional graphics and interactive interfaces that augmented reality promises it will rival the very best science fiction novels. Users will want to have shared experiences in these rich augmented reality scenarios, but surely users will want to restrict who can see their content. It is currently unclear how users of such devices will authenticate one another, particularly in scenarios where access to the Internet is restricted or undesirable. Traditional authentication protocols rely on a trusted authority to bootstrap authentication between two users, but augmented reality content sharing will usually occur in face-to-face scenarios where it will be advantageous for both performance and usability reasons to keep communications and authentication localized. Looks Good To Me (LGTM) is an authentication protocol for augmented reality headsets that leverages the unique hardware and context provided with augmented reality headsets to solve an old problem in a more usable and more secure way. LGTM works over point to point wireless communications so users can authenticate one another in any circumstance and is designed with usability at its core, requiring users to perform only two actions: one to initiate and one to con rm. LGTM allows users to intuitively authenticate one another, using seemingly only each other's faces. Under the hood LGTM uses a combination of facial recognition and wireless localization to ensure secure and extremely simple authentication.


You are cordially invited to attend Bing Wang's Ph.D. Final Defense
Apr. 28, 2016, 1:00-3:00pm, Room NVC T3



Speaker: Bing Wang
Advisor: Prof. Wenjing Lou


Title: Search over Encrypted Data in Cloud Computing

Abstract:


Cloud computing which provides its users computation and storage resources in a pay-per-usage manner has emerged as the most popular computation paradigm in nowadays. Under the new paradigm, users are able to request computation resources dynamically in real-time to accommodate their computation tasks. The elasticity of the flexible resource allocation endows cloud computing services to be able to o er affordable and efficient computation services. However, moving data and applications into the cloud exposes a privacy leakage rick of the user data. As the growing awareness of data privacy, more and more users choose to proactively protect their data in the cloud through encryption. One major problem of data encryption is that it hinders necessary data utilization functions since most of them cannot be applied to encrypted data. The problem could potentially jeopardize the popularity of cloud computing, therefore, achieving efficient data utilization over encrypted data while preserving user data privacy is crucial in cloud computing.

The focus of this dissertation is to design secure and efficient schemes to address essential data utilization functions over encrypted data in cloud computing. To this end, we make contributions in three directions, i.e., to support approximate keyword matching over encrypted data, to design a verifiable and secure variation of the most popular search index - inverted index, and to provide data privacy for the privacy-sensitive data applications in cloud. The first problem that is studied in this dissertation is fuzzy multi-keyword search over encrypted data as fuzzy search is the mostly used and essential data utilization function in our daily life. We propose a novel design that incorporates Bloom Filter and Locality-Sensitive Hashing to fulfill the security and functional requirements of the problem. Secondly, we propose a secure index which is based on the most popular index structure, i.e., the inverted index. Our innovative design provides privacy protection over the secure index, the user query as well as the search pattern and the search result. In addition, users are able to verify the correctness of the search results to ensure the proper computation is performed by the cloud. Finally, we focus ourselves on the privacy-sensitive data application in cloud, i.e., genetic testing over DNA sequences. To provide secure and efficient genetic testing in cloud, we utilize Predicate Encryption and design a bi-linear paring based secure sequence matching scheme to achieve strong privacy guarantee while fulfilling the functionality requirement efficiently. In all of the three research thrusts, we present thorough theoretical security analysis and extensive simulation studies to evaluate the performance of the proposed schemes. The results demonstrate that the proposed schemes can effectively and efficiently address the challenging problems in practice.


You are cordially invited to attend Liang Zhao's Ph.D. Research Defense
Apr. 27, 2016, 9:00-11:00am, Room NVC 320



Speaker: Liang Zhao
Advisor: Prof. Chang-Tien Lu


Title: Spatio-temporal Event Detection and Forecasting in Social Media

Abstract:


Nowadays, knowledge discovery on social media has attracted growing interests. Social media, far more than a communication tool, functions as social sensors for our society. Hundreds of millions of users collectively post millions of tweets every hour, discussing a variety of content ranging from everyday feelings to erent from traditional documents, social media exhibits many other interesting properties: 1) Timeliness of messages: Unlike traditional media that take hours or days to publish, tweets can be posted instantly utilizing portable mobile devices; 2) Ubiquity of social sensors: Tweets reflect the public mood and trends, which could be the determinants of future social events; and 3) Availability of geoinformation: Twitter users provide rich location information in proles, texts, and geo-tags.

This thesis focuses on the development of methods for social media-based spatiotemporal event topics and assumptions. Four methods are proposed, including dynamic query expansion for event detection, a generative framework for event forecasting, multi-task learning for spatiotemporal event forecasting, and deep learning based epidemics modeling for forecasting. For the first method, existing solutions for spatiotemporal event detection are mostly supervised and lack flexibility to handle dynamic keywords of social media. The proposed contributions of our work include: (1) Develop an unsupervised framework; (2) Design a novel dynamic query expansion (DQE) method; and (3) Propose an innovative local modularity spatial scan (LMSS) algorithm.

For the second method, traditional solutions for spatiotemporal event forecasting were mostly designed based on simple assumptions and cannot handle the complexity of our task, such as capture the spatiotemporal context, model mixed-type observations, and utilize prior geographical knowledge. The proposed contributions of our work for this task include: (1) Propose a novel generative model for spatial event algorithm for model parameter inference; and (3) Develop a new sequence likelihood calculation method. For the third method, traditional solutions cannot jointly consider the heterogeneity and effectively handle the dynamics of social media data. The proposed contributions for this work include: (1) Formulate a multi-task learning framework for event forecasting; and (2) Jointly model static and dynamic terms. (3) Develop efficient parameter optimization algorithms. For the last method, existing work on epidemics modeling either cannot ensure the timeliness of disease surveillance, or cannot effectively characterize the underlying epidemics mechanism. The contributions of this work include: Propose a novel integrated framework for computational epidemiology and social media mining (2) Develop a semi-supervised multilayer perceptron (MLP) for mining epidemic features, and (3) Design an online training algorithm.


You are cordially invited to attend Yating Wang's Ph.D. Final Defense
Apr. 25, 2016, 9:30-11:30am, Room NVC T3



Speaker: Yating Wang
Advisor: Prof. Ing-Ray Chen


Title: Trust-Based Service Management for Service-Oriented MANETs and Its Application to Service Composition and Task Assignment with Multi-Objective Goals

Abstract:


With the proliferation of fairly powerful mobile devices and ubiquitous wireless technology, traditional mobile ad hoc networks (MANETs) now migrate into a new era of service-oriented MANETs wherein a node can provide and receive service from other nodes it encounters and interacts with. This dissertation research concerns trust management and its applications for service-oriented MANETs to answer the challenges of MANET environments, including no centralized authority, dynamically changing topology, limited bandwidth and battery power, limited observations, unreliable communication, and the presence of malicious nodes who act to break the system functionality as well as selfish nodes who act to maximize their own gain.

We propose a context-aware trust management model called CATrust for service-oriented ad hoc networks. The novelty of our design lies in the use of logit regression to dynamically estimate trustworthiness of a service provider based on its service behavior patterns in a context environment, treating channel conditions, node status, service payoff, and social disposition as “context” information. We develop a recommendation filtering mechanism to effectively screen out dishonest recommendations even in extremely hostile environments in which the majority recommenders are dishonest. We demonstrate desirable convergence, accuracy, and resiliency properties of CATrust. We also demonstrate that CATrust outperforms contemporary peer-to-peer and Internet of Things trust models in terms of service trust prediction accuracy against collusion recommendation attacks.

We validate the design of trust-based service management based on CATrust with a node-to-service composition and binding MANET application and a node-to-task assignment MANET application.

Requested services in a service-oriented MANET very often must be decomposed into more abstract services and then bound. We formulate this as a multi-objective optimization (MOO) problem to minimize the service cost, while maximizing the quality of service and quality of information in the service a user receives. The MOO problem is an SP-to-service assignment problem. We propose a trust-based service composition and binding algorithm to solve the MOO problem. Our proposed algorithm effectively filters out malicious nodes exhibiting various attack behaviors by penalizing them with trust loss, which ultimately leads to high user satisfaction. Further, our proposed algorithm is efficient with linear runtime complexity while achieving a close-to-optimal solution. We carry out extensive simulation to test the relative performance of the proposed algorithm built on CATrust against non-trust-based and trust-based counterparts built on the well-known Beta Reputation System scheme.

Requested services in a service-oriented MANET also very often must handle dynamically arriving tasks to achieve multiple conflicting objectives. We devise a trust-based heuristic algorithm based on auctioning to solve this node-to-task assignment problem with MOO requirements. Our trust-based heuristic algorithm has a polynomial runtime complexity, rather than an exponential runtime complexity as in existing work, thus allowing dynamic node-to-task assignment to be performed at runtime. Further, our trust-based heuristic algorithm built on CATrust outperforms a non-trust-based counterpart using blacklisting techniques and a counterpart built on Beta Reputation, while performing close to the ideal solution quality with perfect knowledge of node status over a wide range of environmental conditions. We conduct extensive sensitivity analysis of the results with respect to key design parameters and alternative trust protocol designs. We also develop a table-lookup method to apply the best trust model parameter settings upon detection of rapid MANET environment changes to maximize MOO performance.


You are cordially invited to attend Yating Wang's Ph.D. Research Defense
Feb. 12, 2016, 9:30-11:30am, Room NVC 320



Speaker: Yating Wang
Advisor: Prof. Ing-Ray Chen


Title: Trust-Based Service Management for Service-Oriented MANETs and Its Application to Service Composition and Task Assignment with Multi-Objective Goals

Abstract:


With the proliferation of fairly powerful mobile devices and ubiquitous wireless technology, traditional mobile ad hoc networks (MANETs) now migrate into a new era of service-oriented MANETs wherein a node can provide and receive service from other nodes it encounters and interacts with. This dissertation research concerns trust management and its applications for service-oriented MANETs to answer the challenges of MANET environments, including no centralized authority, dynamically changing topology, limited bandwidth and battery power, limited observations, unreliable communication, and the presence of malicious nodes who act to break the system functionality as well as selfish nodes who act to maximize their own gain.

We propose a context-aware trust management model called CATrust for service-oriented ad hoc networks. The novelty of our design lies in the use of logit regression to dynamically estimate trustworthiness of a service provider based on its service behavior patterns in response to rapid MANET environment changes. We develop a recommendation filtering mechanism to effectively screen out dishonest recommendations even in extremely hostile environments in which the majority recommenders are dishonest. We demonstrate desirable convergence, accuracy, and resiliency properties of CATrust. We also demonstrate that CATrust outperforms contemporary peer-to-peer and Internet of Things trust models in terms of service trust prediction accuracy against collusion recommendation attacks.

Requested services in a service-oriented MANET very often must be decomposed into more abstract services and then bound. We formulate this as a multi-objective optimization (MOO) problem to minimize the service cost, while maximizing the quality of service and quality of information in the service a user receives. The MOO problem is an SP-to-service assignment problem. We propose a trust-based service composition and binding algorithm to solve the MOO problem. Our proposed algorithm effectively filters out malicious nodes exhibiting various attack behaviors by penalizing them with loss of reputation, which ultimately leads to high user satisfaction. Further, our proposed algorithm is efficient with linear runtime complexity while achieving a close-to-optimal solution. We carry out extensive simulation to test the relative performance of the proposed algorithm built on CATrust against non-trust-based and trust-based counterparts built on the well-known Beta Reputation scheme.

Requested services in a service-oriented MANET also very often must handle dynamically arriving tasks to achieve multiple conflicting objectives. We devise a trust-based heuristic algorithm based on auctioning to solve this node-to-task assignment problem with MOO requirements. Our trust-based heuristic algorithm has a polynomial runtime complexity, rather than an exponential runtime complexity as in existing work, thus allowing dynamic node-to-task assignment to be performed at runtime. Further, our trust-based heuristic algorithm built on CATrust outperforms a non-trust-based counterpart using blacklisting techniques and a counterpart built on Beta Reputation, while performing close to the ideal solution quality with perfect knowledge of node status over a wide range of environmental conditions. We conduct extensive sensitivity analysis of the results with respect to key design parameters and alternative trust protocol designs. We also develop a table-lookup method to apply the best trust protocol parameter settings upon detection of rapid MANET environment changes to maximize MOO performance.


You are cordially invited to attend Yao Zheng's Ph.D. Preliminary Exam Defense
Date: Feb. 5, 2016, 2-4pm, Room NVC 320



Speaker: Yao Zheng
Advisor: Prof. Wenjing Lou


Title: Privacy Preservation for Cloud-Based Data Sharing and Data Analytic

Abstract:


In my Ph.D. dissertation research, I took a multi-disciplinary approach to address privacy concerns in a cloud environment by combining methods and techniques from cryptography, steganography, distributed computing, and diff erential-privacy. I implemented individual systems that address the four critical aspects of privacy including data privacy, location privacy, association privacy, and identity privacy. In contrast to existing approaches which solely rely on a single measure (e.g., cryptographic means) to safeguard privacy, I posited and demonstrated that, depending on the predictability and sensitivity of the information, privacy protection and service quality can be simultaneously guaranteed using a multi-pronged approach.


You are cordially invited to attend Bing Wang's Ph.D. Research Defense
Jan. 25, 2016, 2:30-4:30pm, Room NVC 320



Speaker: Bing Wang
Advisor: Prof. Wenjing Lou


Title: Protecting Data Privacy in Cloud Computing Applications

Abstract:


Cloud computing which provides its users computation and storage resources in a pay-per-usage manner has emerged as the most popular computation paradigm in nowadays. Under the new paradigm, users are able to request computation resources dynamically in realtime to accommodate their computation tasks. The elasticity of flexible resource allocation endows cloud computing services to be able to off er affordable and efficient computation services. However, moving data and applications into the cloud exposes a privacy leakage rick of the user data. As the growing awareness of data privacy, more and more users choose to proactively protect their data in the cloud through encryption. One major problem of data encryption is that it hinders necessary data utilization functions since most of them cannot be applied to encrypted data. The problem could potentially jeopardize the popularity of cloud computing, therefore, achieving e fficient data utilization over encrypted data while preserving user data privacy is crucial in cloud computing.

The focus of this dissertation is to design secure and e fficient schemes to address essential data utilization functions over encrypted data in cloud computing. To this end, we make contributions in three directions, i.e., the basic data utilization functions, the data storage in cloud, and the privacy-sensitive data applications in cloud. The first problem that is studied in this dissertation is fuzzy multi-keyword search over encrypted data as it is the mostly used and essential data utilization function in our daily life. We propose a novel design that incorporates Bloom Filter and Locality-Sensitive Hashing to fulfi ll the security and functional requirements of the problem. Secondly, we design a dynamic searchable encryption based on Private Set Intersection to address the secure inverted index problem as the existing schemes fail to meet the security or efficiency requirements. In addition, users are able to verify the correctness of the search results in our scheme. Thirdly, we examine the potential data privacy leakage caused by the side-channel attacks in cloud storage with data de-duplication enabled. We propose a game theory based framework to mitigate the privacy leakage efficiently in terms of economic cost for cloud storage providers. Finally, we focus ourselves on the privacy-sensitive data application in cloud, i.e., genetic testing over DNA sequences. To provide secure and efficient genetic testing in cloud, we utilize Predicate Encryption and design a bilinear paring based secure sequence matching scheme to achieve strong privacy guarantee while fulfilling the functionality requirement efficiently. In all of the four research thrusts, we present thorough theoretical security analysis and extensive simulation studies using real-world data sets to evaluate the performance of the proposed schemes. The results demonstrate that the proposed schemes can e ffectively and efficiently address the challenging problems in practice.


You are cordially invited to attend Ning Zhang's Ph.D. Research Defense
Jan. 22, 2016, 2-4pm, Room NVC 320



Speaker: Ning Zhang
Advisor: Prof. Wenjing Lou


Title: Attack and Defense with Hardware Aided Security

Abstract:


Riding on recent advances in computing and networking, our society is now experiencing the evolution into the age of information. While the development of these technologies brings great benefits to our daily lives, the lucrative reward from cyber-crimes have also attracted criminals. Unwanted software are delivered to the victims through various methods such as remote exploitation and social engineering. As computing continues to play an increasing role in the society, security has become a pressing issues. Failures in computing systems could result in loss of infrastructure or human lives, as demonstrated in both research in academia and actual deployment in the wild. With the ever increasing in number of malicious software and new vulnerabilities revealing every day, protecting the heterogeneous computing systems across the Internet has become a daunting task.

My approach to this challenge consists of two direction. The first direction aims to gain better understanding of the inner working of both attacks and defenses in the cyber environment. It was said in Art of War by Sun Tzu, ”If you know your enemies and know yourself, you will not be imperiled in a hundred battles”. I share the same belief that in order to design systems that are resistant to cyber attackers, it is necessary to understand how cyber-attack works. Under this direction, some of my research focus on examining the evolution of cyber-attack. The second direction, on the other hand, attempts to design system using security features offered by the hardware. Hardware supported security offers the opportunity to have secure environment even when the software environment is compromised. Under this direction, my research focus on providing trusted environment in an adversarial setting with the presence of powerful attackers.

The next generation of computing consist powerful elastic cloud and mobile devices inter-connecting via fast reliable networks. Therefore to provide a complete solution, our work spans both the cloud and mobile end point.


You are cordially invited to attend Liang Zhao's Ph.D. Preliminary Exam Defense
Oct. 22, 2015, 10:00-12:00am, Room NVC 320



Speaker: Liang Zhao
Advisor: Prof. Chang-Tien Lu


Title: Spatio-temporal Event Detection and Forecasting in Social Media

Abstract:


Nowadays, knowledge discovery on social media has attracted growing interests. Social media, far more than a communication tool, functions as social sensors for our society. Hundreds of millions of users collectively post millions of tweets every hour, discussing a variety of content ranging from everyday feelings to erent from traditional documents, social media exhibits many other interesting properties: 1) Timeliness of messages: Unlike traditional media that take hours or days to publish, tweets can be posted instantly utilizing portable mobile devices; 2) Ubiquity of social sensors: Tweets re ect the publics mood and trends, which could be the determinants of future social events; and 3) Availability of geo-information: Twitter users provide rich location information in proles, texts, and geotags.

This thesis focuses on the development of methods for social media-based spatiotemporal erent event topics and assumptions. Four methods are proposed, including dynamic query expansion for event detection, a generative framework for event forecasting, multi-task learning for spatiotemporal event forecasting, and deep learning based epidemics modeling for forecasting. For the rst method, existing solutions for spatiotemporal event detection are mostly supervised and lack exibility to handle dynamic keywords of social media. The proposed contributions of our work include: (1) Develop an unsupervised framework; (2) Design a novel dynamic query expansion (DQE) method; and (3) Propose an innovative local modularity spatial scan (LMSS) algorithm.

For the second method, traditional solutions for spatiotemporal event forecasting were mostly designed based on simple assumptions and cannot handle the complexity of our task, such as capture the spatiotemporal context, model mixed-type observations, and utilize prior geographical knowledge. The proposed contributions of our work for this task include: (1) Propose a novel generative model for spatial event ective algorithm for model parameter inference; and (3) Develop a new sequence likelihood calculation method. For the third method, traditional solutions cannot jointly consider the heterogeneity ectively handle the dynamics of social media data. The proposed contributions for this work include: (1) Formulate a multi-task learning framework for event forecasting; and (2) Jointly model static and dynamic terms. (3) Develop efficient parameter optimization algorithms. For the last method, existing work on epidemics modeling either cannot ensure the timeliness of disease surveillance, or cannot ectively characterize the underlying epidemics mechanism. The contributions of this work include: Propose a novel integrated framework for computational epidemiology and social media mining (2) Develop a semi-supervised multilayer perceptron (MLP) for mining epidemic features, and (3) Design an online training algorithm.


You are cordially invited to attend Bing Wang's Ph.D. Preliminary Exam Defense
Sept. 9, 2015, 9:30-11:30am, Room NVC 207



Speaker: Bing Wang
Advisor: Prof. Wenjing Lou


Title: Protecting Data Privacy in Cloud Computing

Abstract:


Traditionally, people store their data and run their computation tasks on their local servers. When entering the cloud computing era, the computation and the storage have been moved into the cloud.

The major benefits of the cloud computing come from these aspect. First of all, it reduce the IT infrastructure cost of the organizations. Instead of purchasing the expensive equipment, business owners can utilize the powerful computation resource of cloud in a pay-as-you-go manner. The pay-per-use approach also optimizes the cost for computation. Secondly, the cloud allows the user to request additional computation resources in real-time to handle the burst of the workload. The feature assures the reliability of the user service under various scenarios. Additionally, the cloud is accessible from anywhere that the Internet is available. Finally, the cloud which is run and managed by professionals provides hassle-free management.

Besides all the benefits, moving the user data into cloud raises a data privacy concern. Because the users lose the control of the physical storage of their data and the potential compromise of the cloud, the user data could be revealed to malicious parties. And the data could be sensitive or classified. Therefore, ensuring the data privacy is an important problem in cloud computing applications.

There are many research areas of data privacy in cloud computing. One of them is computation outsourcing security. The objective is to protect the input data privacy while allowing the computation to be performed in the cloud. Another research area is related to secure data storage. The last one is data service outsourcing security. Search is one of the data service. Our group has done a lot of work on this topic.

My research starts from the searchable encryption schemes. Along this line, I focus on designing more practical scheme that provides popular function that is used in plaintext search and building secure index from the widely adopted index structure, inverted index. Then, because cloud computing is not only a concept but has become a business, we should take economic factors into consideration when designing security schemes in cloud. To that end, we use the side-channel attack in data storage as our problem and we develop a game theoretic framework to model the problem with economic consideration.


You are cordially invited to attend Ning Zhang 's Ph.D. Preliminary Exam Defense
March 27, 2015, 9-11am, NVC 203



Speaker: Ning Zhang
Advisor: Prof. Wenjing Lou


Title: Attack and Defense with Hardware Aided Security

Abstract:


Riding on the benefits of rapid information flow facilitated by the recent advance in computing and networking, our society is now experiencing the evolution into the age of information. While the development of these technologies brings great benefits to our daily lives, the lucrative reward from cyber crimes have also attracted criminals. Unwanted software are delivered to the victims through various methods including remote exploitation, social engineering and etc. As the role of computing continues to couple into the society, security has become a pressing issues. A lost of control on the computing system could very much implies lost of infrastructure or human lives, as demonstrated in both research in academia and actual deployment in the wild. With the ever increasing in number of malicious software and new vulnerabilities revealing everyday, protecting the computing systems across the Internet has become a daunting task.

I believe there are two ways to answer this challenge. The first approach aims at gaining better understand the inner working of both attacks and defenses in the cyber environment. It was said that If you know your enemies and know yourself, you will not be imperiled in a hundred battles. I share the same belief that in order to produce better system that is resistant to cyber attackers, we will have to understand how the attacks work and how it will evolve. With this belief, two of my research work look at how to attack current system. The second approach on the other hand, attempts to tackle the problem by utilizing constructs that is out of reach of software attacks. Hardware supported security offers the opportunity to have secure environment even when the software environment is compromised. From this perspective, we researched how to provide a trusted environment in an adversarial setting with the presence of powerful attackers. We envision that the next generation of computing consist powerful elastic cloud with desk stations and mobile devices inter-connecting via fast reliable networks. Therefore to provide a complete solution, our work spans both the cloud and mobile end point.


You are cordially invited to attend Yating Wang 's Ph.D. Preliminary Exam Defense
Sept 5, 2014, 1-4pm, NVC 320



Speaker: Yating Wang
Advisor: Prof. Ing-Ray Chen


Title: Trust Management for Service-Oriented MANETs and Its Application to Service Composition and Task Assignment with Multi-Objective Goals

Abstract:


With the proliferation of fairly powerful mobile devices and ubiquitous wireless technology, traditional mobile ad hoc networks (MANETs) now migrate into a new era of service oriented MANETs wherein a node can provide and receive service from other nodes it encounters and interacts with. This dissertation research concerns trust management and its applications for service-oriented MANETs to answer the challenges of MANET environments, including no centralized authority, dynamically changing topology, limited bandwidth and battery power, limited observations, unreliable communication, and the presence of malicious nodes who act to break the system functionality as well as selfish nodes who act to maximize their own gain.

This dissertation research has three goals: (1) identifying trust dimensions for service-oriented MANET applications; (2) developing an efficient and efficient trust protocol for service-oriented MANETs; and (3) developing efficient and effective trust-based algorithms for a set of service-oriented MANET applications to achieve multi-objective optimization close to the ideal solution. We develop design principles for achieving these goals. Our overarching principle is the design notion of adaptive control, allowing trust computation, aggregation, propagation, formation (out of multiple trust dimensions) and update decisions to be dynamically adjusted to save energy without compromising trust accuracy and resiliency.

We propose a novel logit regression-based trust model called LogitTrust to dynamically estimate the trust of a node based on how it behaves in response to operational and environment changes. We demonstrate that LogitTrust outperforms traditional approaches based on Bayesian Inference with belief discounting in terms of trust accuracy and resiliency against attacks, while maintaining a low false positive rate. We develop, test, and validate multidimensional trust-based algorithms based on the design notion of application performance optimization to achieve multi-objective optimization for a class of service-oriented MANET applications, including a node-to-service composition and binding problem, a node-to-task assignment problem, and a node-to-group coalition formation problem. We demonstrate that our multi-trust-based algorithms for solving these problems are efficient and effective without compromising solution optimality when compared with non-trust-based solutions, and other trust-based solutions based on Bayesian inference or fuzzy logic.


You are cordially invited to attend Mehmet Saglam's MS Thesis Defense
May 5, 2013, 11:00am-12:30pm, Room NVC 207



Speaker: Mehmet Saglam
Advisor: Prof. Ing-Ray Chen


Title: A Military Planning Methodology for Conducting Cyber Attacks on Power Grid

Abstract:


Power grids are regarded as significant military targets and have been targeted with kinetic attacks in previous military operations. These attacks resulted in significant levels of physical destruction, which, in the long-term, both undermined the success of the operations and caused severe adverse effects on the human terrain. Since power grids have grown as a result of introducing advanced technologies, they have also become more dependent upon cyberspace and are thus exposed to cyber attacks. Since cyber attacks have demonstrated the ability to creating physical/nonphysical effects with surgical precision, they have emerged as a credible option for disrupting power operations for a reasonable duration. However, these types of attacks sometimes require complex coordination with entities from distinct fields for efficient planning; a lack of awareness of the global picture about how to conduct these attacks could result in miscalculations and cause a repeat of the same past failures.

Motivated by this fact, this thesis holistically analyzes the steps involved in conducting cyber attacks on power grids for the purpose of gaining military superiority and provides a comparison for the capabilities, challenges, and opportunities of kinetic and cyber attacks. For the purpose of creating a comprehensive framework for this thesis, the following considerations have been incorporated: the analyses of goals, targets, solutions, and effects of previous military operations; the physical and cyber infrastructures of power grids; and the features, challenges, and opportunities of cyber attacks. To present the findings, this document has adopted a novel military methodology for both the cyber attack analysis and the comparison of the means.


You are cordially invited to attend Qiben Yan's Ph.D Final Defense
June 23, 2014, 10-12am, Room NVC 207



Speaker: Qiben Yan
Advisor: Prof. Wenjing Lou


Title: Toward Security Enhanced Cognitive Radio Networks

Abstract:


ognitive Radio (CR) has been envisioned as a new wireless paradigm to better utilize the spectrum resources, by allowing unlicensed users to opportunistically access the licensed spectrum bands without inducing interference to licensed users. The CR technology has been widely accepted as a solution to the spectrum shortage problem, by improving the spectrum utilization through dynamic spectrum allocations. Recent advances in CR technology have enabled practical broadband CR communications in the TV whitespace. However, the security-related exploration of CR networks is still in its infancy. Without security consideration in mind, various security breaches exploited by sophisticated adversaries could compromise the social welfare of the CR technology.

The focus of this dissertation is to exploit the security vulnerabilities of the state-of-theart CR communication technologies, and to provide detection, mitigation and protection mechanisms to allow security enhanced CR communications. Specifically, we focus on securing two enabling functionalities of a CR network, including spectrum sensing and spectrum opportunity exploitation. Toward securing distributed spectrum sensing, we conduct a comprehensive vulnerability analysis of a distributed concensus-based spectrum sensing algorithm, and propose effective protection mechanisms to thwart advanced attackers who can adaptively adjust their attack strategies by perceiving the surrounding environments. For protecting spectrum opportunity exploitation, we develop a systematic passive monitoring framework, SpecMonitor, based on unsupervised machine learning methods to strategically capture the network traffic, as the basis of detecting anomalous network behaviors. Furthermore, we demonstrate an application of traffic monitoring in facilitating network-wide Peer-to-Peer (P2P) botnet detection, which leads us to design a data-driven system, Peer-Clean, to identify P2P botnets through network flow analysis based on an exploitation of the dynamic group-level behaviors of the bot-infected machines. Finally, deemed as the key enabler of cognitive radio networks, the highly programmable software-defined radio technology has made high-power, wideband and reactive jamming attacks realistic, which imposes a more serious threat to the wireless networks than traditional jamming. By exploiting MIMO technology, we propose jamming resistant communications to turn a non-connectivity scenario into an operational network. In addition, we present thorough security analysis, extensive simulations and testbed evaluations based on real-world implementations. Our results demonstrate that the proposed mechanisms can effectively and efficiently counteract sophisticated yet powerful attacks.

This dissertation contributes to advancing the state-of-the-art security research in modern network design. The discovered security loopholes in the existing networking systems are astounding, which urgently call for a system redesign with a series of thorough security tests or a comprehensive set of effective defending mechanisms in place. The proposed defense mechanisms in this dissertation are designed correspondingly with practical implementations in mind, and evaluated with prototype and testbed designs. We believe these defense mechanisms make an compelling effort to guard and fortify the modern networks.


You are cordially invited to attend Raimundo F. Dos Santos Jr.'s Ph.D Final Defense
May 19, 2014, 10am, Room NVC 320



Speaker: Raimundo F. Dos Santos Jr.
Advisor: Prof. Chang-Tien Lu


Title: Data Analysis in Spatial Contexts

Abstract:


With the growing spread of spatial data, exploratory analysis has gained a considerable amount of attention. Particularly in the fields of Information Retrieval and Data Mining, the integration of data points helps uncover interesting patterns not always visible to the naked eye. Social networks often link entities that share places and activities; marketing tools target users based on behavior and preferences; and medical technology combines symptoms to categorize diseases. Many of the current approaches in this field of research rely on syntactic heuristics, which are good for comparisons, but less than ideal for inferences. Others apply semantic methods helpful in drawing extended conclusions, but which fail to incorporate syntacts appropriately. This research focuses on spatial data analysis that incorporates both syntactic and semantic methods.

From a functional perspective, any spatial object can be investigated at the entity or at the attribute levels. The former attempts to predict how two entities are alike, corresponding to a semantic view; the latter makes no assumptions about the entities as a whole, but rather, observes if any of their attributes are similar, which leans toward a syntactic view. Existing research examines several aspects of entities and their attributes: shared relationships among objects, matches versus mismatches of values, distances among parents and children, and brute-force comparison of er from the pitfalls of disparate data, often missing true relationships, failing to deal with inexact vocabularies, ignoring missing values, and poorly handling multiple attributes. In addition, the vast majority does not consider the spatial aspects of the data.

This research combines semantic and syntactic techniques of data analysis in spatial contexts. erent similarity measures among spatial entities as well as among sequences of entities. They are able to identify relationships that are not explicitly written down. Major contributions of this research include (1) a framework that computes a numerical entity similarity, denoted a semantic footprint, composed of spatial, dimensional, and ontological facets; (2) a semantic approach that translates categorical data into a numerical similarity, which permits ranking and ordering; (3) an extensive study of GML as a representative spatial structure of how semantic analysis methods are influenced by its approaches to storage, querying, and parsing; (4) a method to find spatial regions of high entity density based on a clustering coe cient; (5) a ranking strategy based on connectivity strength erentiates important relationships from less relevant ones; (6) a distance measure between entity sequences that quantifies the most related streams of information; (7) three distance-based measures (one probabilistic, one based on spatial influence, and one spatio-logical) that quantifies the interactions among entities and events.


You are cordially invited to attend Hamid Al-Hamadi's Ph.D Final Defense
April 2, 2014, 2pm-4pm, Room NVC 320



Speaker: Hamid Al-Hamadi
Advisor: Prof. Ing-Ray Chen


Title: Dynamic Redundancy Management of Multisource Multipath Routing Integrated with Voting-based Intrusion Detection in Wireless Sensor Networks

Abstract:


Wireless sensor networks (WSNs) are frequently deployed unattended and can be easily captured or compromised. Once compromised, intrusion preven-tion methods such as encryption can no longer provide any protection, as a com-promised node is considered a legitimate node and possesses the secret key for decryption. Compromised nodes are essentially inside attackers and can perform various attacks to break the functionality of the system. Thus, for safety-critical WSNs, intrusion detection techniques must be used to detect and remove inside attackers and fault tolerance techniques must be used to tolerate inside attackers to prevent security failure.

In this dissertation research, we develop a class of dynamic redundancy man-agement algorithms for redundancy management of multisource multipath rout-ing for fault and intrusion tolerance, and majority voting for intrusion detection, with the goal of maximizing the WSN lifetime while satisfying application quali-ty-of-service and security requirements, for base station based WSNs, homoge-neous clustered WSNs, and heterogeneous clustered WSNs. By means of a novel model-based analysis methodology based on probability theory, we model the tradeoff between energy consumption vs. reliability, timeliness and security gain, and identify the optimal multisource multipath redundancy level and intrusion detection settings for maximizing the lifetime of the WSN while satisfying application quality-of-service requirements. A main contribution of our research dissertation is that our dynamic redundancy management protocol design addresses the issues of "how many paths to use" and "what paths to use" in multisource multipath routing for intrusion tolerance. Another contribution is that we take an integrated approach combining intrusion detection and tolerance in the protocol design to address the issue of "how much intrusion detection is enough" to prevent security failure and prolong the WSN lifetime time.

We demonstrate resiliency of our dynamic redundancy management protocol design for intrusion detection and tolerance against sophisticated attacker behaviors, including selective and random capture, as well as persistent, random, opportunistic and insidious attacks, by model-based performance analysis with results supported by extensive simulation based on ns3.







You are cordially invited to attend Raimundo F. Dos Santos Jr.'s Ph.D. Research Defense
Feb. 21, 2014, 10am, Room NVC 320



Speaker: Raimundo F. Dos Santos Jr.
Advisor: Prof. Chang-Tien Lu


Title: Data Analysis in Spatial Contexts

Abstract:


With the growing spread of spatial data, exploratory analysis has gained a considerable amount of attention. Particularly in the fields of Information Retrieval and Data Mining, the integration of data points helps uncover interesting patterns not always visible to the naked eye. Social networks often link entities that share places and activities; marketing tools target users based on behavior and preferences; and medical technology combines symptoms to categorize diseases. Many of the current approaches in this field of research rely on syntactic heuristics, which are good for comparisons, but less than ideal for inferences. Others apply semantic methods helpful in drawing extended conclusions, but which fail to incorporate syntacts appropriately. This research focuses on spatial data analysis that incorporates both syntactic and semantic methods.

From a functional perspective, any spatial object can be investigated at the entity or at the attribute levels. The former attempts to predict how two entities are alike, corresponding to a semantic view; the latter makes no assumptions about the entities as a whole, but rather, observes if any of their attributes are similar, which leans toward a syntactic view. Existing research examines several aspects of entities and their attributes: shared relationships among objects, matches versus mismatches of values, distances among parents and children, and brute-force comparison of er from the pitfalls of disparate data, often missing true relationships, failing to deal with inexact vocabularies, ignoring missing values, and poorly handling multiple attributes. In addition, the vast majority does not consider the spatial aspects of the data.

This research combines semantic and syntactic techniques of data analysis in spatial contexts. erent similarity measures among spatial entities as well as among sequences of entities. They are able to identify relationships that are not explicitly written down. Major contributions of this research include (1) a framework that computes a numerical entity similarity, denoted a semantic footprint, composed of spatial, dimensional, and ontological facets; (2) a semantic approach that translates categorical data into a numerical similarity, which permits ranking and ordering; (3) an extensive study of GML as a representative spatial structure of how semantic analysis methods are influenced by its approaches to storage, querying, and parsing; (4) a method to find spatial regions of high entity density based on a clustering coe cient; (5) a ranking strategy based on connectivity strength erentiates important relationships from less relevant ones; (6) a distance measure between entity sequences that quantifies the most related streams of information; (7) three distance-based measures (one probabilistic, one based on spatial influence, and one spatio-logical) that quantifies the interactions among entities and events.







You are cordially invited to attend Qiben Yan's Ph.D Research Defense
Jan. 23, 2014, 10-12am, Room NVC 351



Speaker: Qiben Yan
Advisor: Prof. Wenjing Lou


Title: Toward Security Enhanced Cognitive Radio Networks

Abstract:


Cognitive Radio (CR) has been envisioned as a new wireless paradigm to better utilize the spectrum resources, by allowing unlicensed users to opportunistically access the licensed spectrum bands without inducing interference to licensed users. The CR technology has been widely accepted as a solution to the spectrum shortage problem, by improving the spectrum utilization through dynamic spectrum allocations. Recent advances in CR technology have enabled practical broadband CR communications in the TV whitespace. However, the security-related exploration of CR networks is still in its infancy. Without security consideration in mind, various security breaches exploited by sophisticated adversaries could compromise the social welfare of the CR technology.

The focus of this dissertation is to exploit the security vulnerabilities of the state-of-theart CR communication technologies, and to provide detection, mitigation and protection mechanisms to allow security enhanced CR communications. Specifically, we focus on securing two enabling functionalities of a CR network, including spectrum sensing and spectrum opportunity exploitation. Toward securing distributed spectrum sensing, we conduct a comprehensive vulnerability analysis of a distributed concensus-based spectrum sensing algorithm, and propose effective protection mechanisms to thwart advanced attackers who can adaptively adjust their attack strategies by perceiving the surrounding environments. For protecting spectrum opportunity exploitation, we develop a systematic passive monitoring framework, SpecMonitor, based on unsupervised machine learning methods to strategically capture the network traffic, as the basis of detecting anomalous network behaviors. Furthermore, we demonstrate an application of traffic monitoring in facilitating network-wide Peer-to-Peer (P2P) botnet detection, which leads us to design a data-driven system, Peer-Clean, to identify P2P botnets through network flow analysis based on an exploitation of the dynamic group-level behaviors of the bot-infected machines. Finally, deemed as the key enabler of cognitive radio networks, the highly programmable software-defined radio technology has made high-power, wideband and reactive jamming attacks realistic, which imposes a more serious threat to the wireless networks than traditional jamming. By exploiting MIMO technology, we propose jamming resistant communications to turn a non-connectivity scenario into an operational network. In addition, we present thorough security analysis, extensive simulations and testbed evaluations based on real-world implementations. Our results demonstrate that the proposed mechanisms can effectively and efficiently counteract sophisticated yet powerful attacks.

This dissertation contributes to advancing the state-of-the-art security research in modern network design. The discovered security loopholes in the existing networking systems are astounding, which urgently call for a system redesign with a series of thorough security tests or a comprehensive set of effective defending mechanisms in place. The proposed defense mechanisms in this dissertation are designed correspondingly with practical implementations in mind, and evaluated with prototype and testbed designs. We believe these defense mechanisms make an compelling effort to guard and fortify the modern networks.







You are cordially invited to attend Hamid Al-Hamadi's Ph.D Research Defense
Nov. 19, 2013, 10am-12pm, Room NVC 320



Speaker: Hamid Al-Hamadi
Advisor: Prof. Ing-Ray Chen


Title: Dynamic Redundancy Management of Multisource Multipath Routing Integrated with Voting-based Intrusion Detection in Wireless Sensor Networks

Abstract:


Wireless sensor networks (WSNs) are frequently deployed unattended and can be easily captured or compromised. Once compromised, intrusion prevention methods such as encryption can no longer provide any protection, as a compromised node is considered a legitimate node and possesses the secret key for decryption. Compromised nodes are essentially inside attackers and can perform various attacks to break the functionality of the system. Thus, for safety-critical WSNs, intrusion detection techniques must be used to detect and remove inside attackers and fault tolerance techniques must be used to tolerate inside attackers to prevent security failure.

In this dissertation research, we develop a class of dynamic redundancy man-agement algorithms for redundancy management of multisource multipath routing for fault and intrusion tolerance, and majority voting for intrusion detection, with the goal of maximizing the WSN lifetime while satisfying application quality-of-service and security requirements, for base station based WSNs, homogeneous clustered WSNs, and heterogeneous clustered WSNs. By means of a novel model-based analysis methodology based on probability theory, we model the tradeoff between energy consumption vs. reliability, timeliness and security gain, and identify the optimal multisource multipath redundancy level and intrusion detection settings for maximizing the lifetime of the WSN while satisfying application quality-of-service requirements. A main contribution of our research dissertation is that our dynamic redundancy management protocol design addresses the issues of how many paths to use and what paths to use in multisource multipath routing for intrusion tolerance. Another contribution is that we take an integrated approach combining intrusion detection and tolerance in the protocol design to address the issue of how much instruction detection is enough to prevent security failure and prolong the WSN lifetime time.

We demonstrate resiliency of our dynamic redundancy management protocol design for intrusion detection and tolerance against sophisticated attacker behaviors, including selective and random capture, as well as persistent and random attacks, by model-based performance analysis validated with extensive simulation based on ns3. We also demonstrate the validity of our design by a compara-tive performance analysis with existing multipath routing protocols through extensive simulation.







You are cordially invited to attend Reghu Anguswamy's Ph.D Final Defense
June 13, 2013, 10:30am-12:30pm, Room NVC T3



Speaker: Reghu Anguswamy
Advisor: Prof. Bill Frakes


Title: Factors Affecting the Design and Use of Reusable Components

Abstract:


Designing software components for future reuse has been an important area in software engineering. A software system developed with reusable components follows a 'with' reuse process while a component designed to be reused in other systems follows a 'for' reuse process. This dissertation explores the factors affecting design for reuse and design with reusable components.

Design for reuse: In this thesis, the first study was conducted analyzing one-use and equivalent reusable components for the overhead in terms of component size, effort required, number of parameters, and productivity. Reusable components were significantly larger than their equivalent one-use components and had significantly more parameters. The effort required for the reusable components was higher than for one-use components. The productivity of the developers was significantly lower for the reusable components compared to the one-use components. Also, during the development of reusable components, the subjects spent more time on writing code than designing the components, but not significantly so. A ranking of the design principles by frequency of use is also reported. A content analysis performed on the feedback is also reported and the reasons for using and not using the reuse design principles are identified. A correlation analysis showed that the reuse design principles were, in general, used independently of each other.

Design with reuse: Through another empirical study, the effect of the size of a component and the reuse design principles used in building the component on the ease of reuse were analyzed. It was observed that the higher the complexity the lower the ease of reuse, but the correlation is not significant. When considered independently, four of the reuse design principles: well-defined interface, clarity and understandability, generality, and separate concepts from content significantly increased the ease of reuse while commonality and variability analysis significantly decreased the ease of reuse, and documentation did not have a significant impact on the ease of reuse. Experience in the programming language had no significant relationship with the reusability of components. Experience in software engineering and software reuse showed a relationship with reusability but the effect size was small. Testing components before integrating them into a system was found to have no relationship with the reusability of components. A content analysis of the feedback is presented identifying the challenges of components that were not easy to reuse. Features that make a component easily reusable were also identified. The Mahalanobis-Taguchi Strategy (MTS) was employed to develop a model based on Mahalanobis Distance to identify the factors that can detect if a component is easy to reuse or not. The identified factors within the model are: size of a component, a set of reuse design principles (well-defined interface, clarity and understandability, commonality and variability analysis, and generality), and component testing.







You are cordially invited to attend Qiben Yan's PhD Preliminary Exam defense
May 28, 2013, 10-12am, Room NVC 351



Speaker: Qiben Yan
Advisor: Prof. Wenjing Lou


Title: Toward Security Enhanced Cognitive Radio Networks

Abstract:


Cognitive Radio (CR) has been envisioned as a new wireless paradigm to better utilize the spectrum resources, by allowing unlicensed users to opportunistically access the licensed spectrum bands without inducing interference to licensed users. CR technology has been a widely-accepted solution to spectrum shortage problem, by improving the spectrum utilization through dynamic spectrum allocations. Recent advances in CR technology have enabled practical broadband CR communications in the TV whitespace. However, the security related exploration of CR networks is still in its infancy. Without security consideration in mind, various security breaches exploited by sophisticated adversaries, could compromise the social welfare of CR technologies.

The focus of this dissertation is to exploit the security vulnerabilities of state-of-the-art CR communication technologies, and provide detection, mitigation and protection mechanisms to allow security enhanced CR communications. Specifically, we focus on securing two enabling functionalities of a CR network, including spectrum sensing and spectrum opportunity exploitation. Toward securing distributed spectrum sensing, we conduct a comprehensive vulnerability analysis of distributed concensus-based spectrum sensing algorithm, and propose effective protection mechanisms to thwart advanced attackers who can adaptively adjust their attack strategies by perceiving the surrounding environments. For protecting spectrum opportunity exploitation, we develop a systematic passive monitoring framework based on unsupervised machine learning methods to strategically capture the network traffic, as the basis of detecting anomalous network behaviors. Furthermore, we demonstrate the utility of traffic monitoring for facilitating real-world Peer-to-Peer botnet detection, which results in a mechanism design of PeerClean, a fully automated method for detecting P2P botnets through network flow analysis, based on the exploitation of dynamic group-level behaviors of bot-infected machines. Finally, the highly programmable software-defined radio technology has made high-power, wideband and reactive jamming attacks realistic, which imposes a more serious threat to the wireless networks than traditional jamming. By exploiting MIMO technology for anti-jamming communication, we are able to turn a non-connectivity scenario into an operational network. Using security analysis, extensive simulations and testbed evaluations based on real-world implementation, we show that the proposed mechanisms can effectively and efficiently counteract sophisticated yet powerful attacks.







You are cordially invited to attend Xutong Liu's Ph.D Final Defense
May 7, 2013, 11am-1pm, Room NVC 351



Speaker: Xutong Liu
Advisor: Prof. Chang-Tien Lu


Title: Prediction and Anomaly Detection Techniques for Spatial Datasets

Abstract:


With increasing public sensitivity and concern on environmental issues, huge amounts of spatial data have been collected from location based social network applications to scientic data. This has encouraged formation of large spatial datasets and generated considerable interests for identifying novel and meaningful patterns. Allowing correlated observations weakens the usual statistical assumption of independent observations, and complicates the spatial analysis. This research focuses on the construction of ecient and ective approaches for three main mining tasks, including spatial outlier detection, robust inference for spatial dataset, and spatial prediction for large multivariate non-Gaussian data.

Spatial outlier analysis, which aims at detecting abnormal objects in spatial contexts, can help extract important knowledge in many applications. There exist the well-known masking and swamping problems in most approaches, which can't still satisfy certain requirements aroused recently. This research focuses on development of spatial outlier detection techniques for three aspects, including spatial numerical outlier detection, spatial categorical outlier detection and identication of the number of spatial numerical outliers.

First, this report introduces Random Walk based approaches to identify spatial numerical outliers. A Bipartite and an Exhaustive Combination weighted graphs are modeled based on spatial and/or non-spatial attributes, and then Random walk techniques are performed on the graphs to compute the relevance among objects. The objects with lower relevance are recognized as outliers. Second, an entropy-based method is proposed to estimate the optimum number of outliers. According to the entropy theory, we expect that, by incrementally removing outliers, the entropy value will decrease sharply, and reach a stable state when all the outliers have been removed. Finally, this research designs several Pair Correlation Function based methods to detect spatial categorical outliers for both single and multiple attribute data. Within them, Pair Correlation Ratio(PCR) is dened and estimated for each pair of categorical combinations based on their co-occurrence frequency at dirent spatial distances. The observations with the lower PCRs are diagnosed as potential SCOs.

Spatial kriging is a widely used predictive model whose predictive accuracy could be signicantly compromised if the observations are contaminated by outliers. Also, due to spatial heterogeneity, observations are often dirent types. The prediction of multivariate spatial processes plays an important role when there are cross-spatial dependencies between multiple responses. In addition, given the large volume of spatial data, it is computationally challenging. These raise three research topics: 1) robust prediction for spatial data sets; 2) prediction of multivariate spatial observations; and 3) efficient processing for large data sets.

First, increasing the robustness of spatial kriging model can be systematically addressed by integrating heavy tailed distributions. However, it is analytically intractable inference. Here, we presents a novel Robust and reduced Rank spatial kriging Model (R3-SKM), which is resilient to the inuences of outliers and allows for fast spatial inference. Second, this research introduces a exible hierarchical Bayesian framework that permits the simultaneous modeling of mixed type variable. Specically, the mixed-type attributes are mapped to latent numerical random variables that are multivariate Gaussian in nature. Finally, the knotbased techniques is utilized to model the predictive process as a reduced rank spatial process, which projects the process realizations of the spatial model to a lower dimensional subspace. This projection signicantly reduces the computational cost.







You are cordially invited to attend Fenye Bao's Ph.D Final Defense
May 6, 2013, 5-7pm, Room NVC 351



Speaker: Fenye Bao
Advisor: Prof. Ing-Ray Chen


Title: Dynamic Trust Management for Mobile Networks

Abstract:


Trust management in mobile networks is challenging due to dynamically changing network environments and the lack of a centralized trusted authority. In this dissertation research, we design and validate a class of dynamic trust management protocols for mobile networks, and demonstrate the utility of dynamic trust management with trust-based applications. Unlike existing work, we consider social trust derived from social networks in addition to traditional quality-ofservice (QoS) trust derived from communication networks to obtain a composite trust metric as a basis for evaluating trust of nodes in mobile network applications. Untreated in the literature, we design and validate trust composition, aggregation, propagation, and formation protocols for dynamic trust management that can learn from past experiences and adapt to changing environment conditions to maximize application performance and enhance operation agility. Furthermore, we propose, explore and validate the design concept of applicationlevel trust optimization in response to changing conditions to maximize application performance or best satisfy application requirements. We provide formal proof for the convergence, accuracy, and resiliency properties of our trust management protocols. To achieve the goals of identifying the best trust protocol setting and optimizing the use of trust for trust-based applications, we develop a novel model-based analysis methodology with simulation validation for analyzing and validating our dynamic trust management protocol design.

The dissertation research provides new understanding of dynamic trust management for mobile wireless networks. We gain insight on the best trust composition and trust formation out of social and QoS trust components, as well as the best trust aggregation and propagation protocols for optimizing application performance. We gain insight on how a modeling and analysis tool can be built, allowing trust composition, aggregation, propagation, and formation designs to be incorporated, tested and validated. We demonstrate the utility of dynamic trust management protocol for mobile networks including mobile ad-hoc networks, delay tolerant networks, wireless sensor networks, and Internet of things systems with practical applications including misbehaving node detection, trust-based survivability management, trust-based secure routing, and trust-based service composition. Through model-based analysis with simulation validation, we show that our dynamic trust management based protocols outperform non-trustbased and Bayesian trust-based protocols in the presence of malicious, erroneous, partly trusted, uncertain and incomplete information, and are resilient to trust related attacks.







You are cordially invited to attend Yang Chen's MS Thesis Defense
May 1, 2013, 10:30-12:00am, Room NVC 313



Speaker: Yang Chen
Advisor: Prof. Charles Clancy


Title: Robust Prediction of Large Spatio-Temporal Datasets

Abstract:


This thesis describes a robust and efficient design of Student-t based Robust Spatio-Temporal Prediction, namely, St-RSTP, to provide estimation based on observations over spatiotemporal neighbors. It is crucial to many applications in geographical information systems, medical imaging, urban planning, economy study, and climate forecasting. The proposed St-RSTP is more resilient to outliers or other small departures from model assumptions than its ancestor, the Spatio-Temporal Random Effects (STRE) model. STRE is a statistical model with linear order complexity for processing large scale spatiotemporal data.

However, it has been shown sensitive to outliers or anomaly observations. In our design, the St-RSTP model assumes that the measurement error follows Students t-distribution, instead of a traditional Gaussian distribution. To handle the analytical intractable inference of Students t model, we propose an approximate inference algorithm in the framework of Expectation Propagation (EP). Extensive experimental evaluations based on both simulation and real-life data sets demonstrated the robustness and the efficiency of our Student-t prediction model compared with the STRE model.







You are cordially invited to attend Alexander K. Buchholz's MS Thesis Defense
April 29, 2013, 2-4pm, Room NVC 351



Speaker: Alexander K. Buchholz
Advisor: Prof. Wenjing Lou


Title: Dual Path PKI for Secure Aircraft Data Communication

Abstract:


Through application of modern technology, aviation systems are becoming more automated and are relying less on antiquated air traffic control (ATC) voice systems. Aircraft are now able to wirelessly broadcast and receive identity and location information using transponder technology. This helps reduce controller workload and allows the aircraft to take more responsibility for maintaining safe separation. However, these systems lack source authentication methods or the ability to check the integrity of message content. This opens the door for hackers to potentially create fraudulent messages or manipulate message content.

This thesis presents a solution to handling many of the potential security issues in aircraft data communication. This is accomplished through the implementation of a Dual Path PKI (DPP) design which includes a novel approach to handling certificate revocation through session certificates. DPP defines two authentication protocols, one between aircraft and another between aircraft and ATC, to achieve source authentication. Digital signature technology is utilized to achieve message content and source integrity as well as enable bootstrapping DPP into current ATC systems. DPP employs cutting-edge elliptic curve cryptography (ECC) algorithms to increase performance and reduce overhead.

It is found that the DPP design successfully mitigates several of the cyber security concerns in aircraft and ATC data communications. An implementation of the design shows that anticipated ATC systems can accommodate the additional processing power and bandwidth required by DPP to successfully achieve system integrity and security.







You are cordially invited to attend Robert Mitchell's Ph.D Final Defense
April 5, 2013, 10am, Room NVC 401



Speaker: Robert Mitchell
Advisor: Prof. Ing-Ray Chen


Title: Design and Analysis of Intrusion Detection Protocols in Cyber Physical Systems

Abstract:


In this dissertation research we aim to design and validate intrusion detection system (IDS) protocols for a cyber physical system (CPS) comprising sensors, actuators, control units, and physical objects for controlling and protecting physical infrastructures.

The design part includes host IDS, system IDS and IDS response designs. The validation part includes a novel model-based analysis methodology with simulation validation. Our objective is to maximize the CPS reliability or lifetime in the presence of malicious nodes performing attacks which can cause security failures. Our host IDS design results in a lightweight, accurate, autonomous and adaptive protocol that runs on every node in the CPS to detect misbehavior of neighbor nodes based on state-based behavior specifications. Our system IDS design results in a robust and resilient protocol that can cope with malicious, erroneous, partly trusted, uncertain and incomplete information in a CPS. Our IDS response design results in a highly adaptive and dynamic control protocol that can adjust detection strength in response to environment changes in attacker strength and behavior. The end result is an energy-aware and adaptive IDS that can maximize the CPS lifetime in the presence of malicious attacks, as well as malicious, erroneous, partly trusted, uncertain and incomplete information.

We develop a probability model based on stochastic Petri nets to describe the behavior of a CPS incorporating our proposed intrusion detection and response designs, subject to attacks by malicious nodes exhibiting a range of attacker behaviors, including reckless, random, insidious and opportunistic attacker models. We identify optimal intrusion detection settings under which the CPS reliability or lifetime is maximized for each attacker model. Adaptive control for maximizing IDS performance is achieved by dynamically adjusting detection and response strength in response to attacker strength and behavior detected at runtime. We conduct extensive analysis of our designs with four case studies, namely, a mobile group CPS, a medical CPS, a smart grid CPS and an unmanned aircraft CPS. The results show that our adaptive intrusion and response designs operating at optimizing conditions significantly outperform existing anomaly-based IDS techniques for CPSs.







You are cordially invited to attend Xutong Liu's Ph.D Research Defense
Jan. 24, 2013, 10-12pm, Room NVC T3



Speaker: Xutong Liu
Advisor: Prof. Chang-Tien Lu


Title: Prediction and Anomaly Detection Techniques for Spatial Datasets

Abstract:


With increasing public sensitivity and concern on environmental issues, huge amounts of spatial data have been collected from location based social network applications to scientic data. This has encouraged formation of large spatial datasets and generated considerable interests for identifying novel and meaningful patterns. Allowing correlated observations weakens the usual statistical assumption of independent observations, and complicates the spatial analysis. This research focuses on the construction of ecient and ective approaches for three main mining tasks, including spatial outlier detection, robust inference for spatial dataset, and spatial prediction for large multivariate non-Gaussian data.

Spatial outlier analysis, which aims at detecting abnormal objects in spatial contexts, can help extract important knowledge in many applications. There exist the well-known masking and swamping problems in most approaches, which can't still satisfy certain requirements aroused recently. This research focuses on development of spatial outlier detection techniques for three aspects, including spatial numerical outlier detection, spatial categorical outlier detection and identication of the number of spatial numerical outliers.

First, this report introduces Random Walk based approaches to identify spatial numerical outliers. A Bipartite and an Exhaustive Combination weighted graphs are modeled based on spatial and/or non-spatial attributes, and then Random walk techniques are performed on the graphs to compute the relevance among objects. The objects with lower relevance are recognized as outliers. Second, an entropy-based method is proposed to estimate the optimum number of outliers. According to the entropy theory, we expect that, by incrementally removing outliers, the entropy value will decrease sharply, and reach a stable state when all the outliers have been removed. Finally, this research designs several Pair Correlation Function based methods to detect spatial categorical outliers for both single and multiple attribute data. Within them, Pair Correlation Ratio(PCR) is dened and estimated for each pair of categorical combinations based on their co-occurrence frequency at dirent spatial distances. The observations with the lower PCRs are diagnosed as potential SCOs.

Spatial kriging is a widely used predictive model whose predictive accuracy could be signicantly compromised if the observations are contaminated by outliers. Also, due to spatial heterogeneity, observations are often dirent types. The prediction of multivariate spatial processes plays an important role when there are cross-spatial dependencies between multiple responses. In addition, given the large volume of spatial data, it is computationally challenging. These raise three research topics: 1) robust prediction for spatial data sets; 2) prediction of multivariate spatial observations; and 3) efficient processing for large data sets.

First, increasing the robustness of spatial kriging model can be systematically addressed by integrating heavy tailed distributions. However, it is analytically intractable inference. Here, we presents a novel Robust and reduced Rank spatial kriging Model (R3-SKM), which is resilient to the inuences of outliers and allows for fast spatial inference. Second, this research introduces a exible hierarchical Bayesian framework that permits the simultaneous modeling of mixed type variable. Specically, the mixed-type attributes are mapped to latent numerical random variables that are multivariate Gaussian in nature. Finally, the knotbased techniques is utilized to model the predictive process as a reduced rank spatial process, which projects the process realizations of the spatial model to a lower dimensional subspace. This projection signicantly reduces the computational cost.







You are cordially invited to attend Gregory P. Bilodeau's MS Thesis Defense
Jan. 2, 2013, 2-3pm, Room NVC 325



Speaker: Gregory P. Bilodeau
Advisor: Prof. Csaba Egyhazy


Title: Automated Seed Point Selection in Confocal Image Stacks Of Neuron Cells

Abstract:


Research into neurological disease and function requires an understanding of how neuron cells connect and interact V a function of their shape, or morphology. To determine the morphology of a neuron, researchers first obtain a series of images of the neuron through a microscope taken at different focal depths called an image stack. This stack is then used to manually trace the neurons 3-dimensional structure, usually using a computer program to manipulate the stack images and track the points created by the researcher.

Although the technology necessary to capture these image stacks has advanced in terms of speed and cost, this process of reconstructing the morphology of neuron cells from image stacks remains a manual, time-consuming and subjective process sometimes taking weeks to months to complete. Fully automated systems capable of constructing these 3-dimensional models with little to no input from the researcher would greatly speed up the research process. Previous attempts to solve this problem have utilized many different techniques, and a recent competition called the DAIDEM Challenge sought to compare different approaches in a competitive manner in order to determine the most effective approaches and spur further innovation in the field.

Many algorithms require the user to supply seed points, specific coordinates that represent a known area of the morphology from which the algorithm can begin a search. Still others need to examine training stacks and reconstructions in order to calibrate themselves, or require other slow IO functions that greatly reduce their performance speed. Proper selection of seed points is a critical task, as many algorithms utilize the local image information around the seeds to direct the tracing process, and knowledge of the global image space is useful in fine tuning the filters and kernels used to produce a successful tracing. I propose an automated method of seed point selection that produces both a comprehensive overview of the global image environment and a set of high-probability seed points while performing faster than conventional means.








You are cordially invited to attend Reghu Anguswamy's Ph.D. Research Defense
Dec. 14, 2012, 11am, Room NVC 221



Speaker: Reghu Anguswamy
Advisor: Prof. Bill Frakes


Title: A Study of Factors Affecting the Design and Use of Reusable Components

Abstract:


Design for reuse: In this thesis, the first study was conducted analyzing one-use and equivalent reusable components for the overhead in terms of component size, effort required, number of parameters, and productivity. Reusable components were significantly larger than their equivalent one-use components and had significantly more parameters. The effort required for the reusable components was higher than for one-use components. The productivity of the developers was significantly lower for the reusable components compared to the one-use components. Also, during the development of reusable components, the subjects spent more time on writing code than designing the components, but not significantly so. A ranking of the design principles by frequency of use is also reported. A content analysis performed on the feedback is also reported and the reasons for using and not using the reuse design principles are identified. A correlation analysis shows that the reuse design principles were, in general, used independently of each other.

Design with reuse: Through another empirical study, the effect of the size of a component and the reuse design principles used in building the component on the ease of reuse were analyzed. It was observed that the higher the complexity the lower the ease of reuse, but the correlation is not significant. When considered independently, four of the reuse design principles: well-defined interface, clarity and understandability, generality, and separate concepts from content significantly increased the ease of reuse while and commonality and variability analysis significantly decreased the ease of reuse, and documentation did not have a significant impact on the ease of reuse. Experience in the programming language had no relationship with the reusability of components. Experience in software engineering and software reuse showed a relationship with reusability but the effect size was not significant. Testing components before integrating them into a system was also found to have no relationship with the reusability of components. A content analysis of the feedback is presented identifying the challenges of components that were not easy to reuse. Features that make a component easily reusable are also identified. The Mahalanobis-Taguchi Strategy (MTS) was employed to develop a model based on Mahalanobis Distance to identify the factors that can detect if a component is easy to reuse or not. The identified factors within the model are: size of a component, a set of reuse design principles (well-defined interface, clarity and understandability, commonality and variability analysis, and generality), and component testing.








You are cordially invited to attend Hamid Al-Hamadi's Ph.D. Preliminary Exam Defense
Dec. 7, 2012, 2pm, Room NVC 401



Speaker: Hamid Al-Hamadi
Advisor: Prof. Ing-Ray Chen


Title: Dynamic Redundancy Management of Multisource Multipath Routing Integrated with Voting-based Intrusion Detection in Wireless Sensor Networks

Abstract:


Wireless sensor networks (WSNs) are frequently deployed unattended and can be easily captured or compromised. Once compromised, intrusion preven-tion methods such as encryption can no longer provide any protection, as a com-promised node is considered a legitimate node and possesses the secret key for decryption. Compromised nodes are essentially inside attackers and can perform various attacks to break the functionality of the system. Thus, for safety-critical WSNs, intrusion detection techniques must be used to detect and remove inside attackers and fault tolerance techniques must be used to tolerate inside attackers to prevent security failure.

In this dissertation research, we develop a class of dynamic redundancy man-agement algorithms for redundancy management of multisource multipath rout-ing for fault and intrusion tolerance, and majority voting for intrusion detection, with the goal of maximizing the WSN lifetime while satisfying application quali-ty-of-service and security requirements, for base station based WSNs, homoge-neous clustered WSNs, and heterogeneous clustered WSNs. By means of a novel model-based analysis methodology based on probability theory, we model the tradeoff between energy consumption vs. reliability, timeliness and security gain, and identify the optimal multisource multipath redundancy level and intrusion detection settings for maximizing the lifetime of the WSN while satisfying appli-cation quality-of-service requirements. A main contribution of our research dis-sertation is that our dynamic redundancy management protocol design address-es the issues of how many paths to use and what paths to use in multisource multipath routing for intrusion tolerance. Another contribution is that we take an integrated approach combining intrusion detection and tolerance in the protocol design to address the issue of how much instruction detection is enough to prevent security failure and prolong the WSN lifetime time.

We demonstrate resiliency of our dynamic redundancy management protocol design for intrusion detection and tolerance against sophisticated attacker behav-iors, including selective and random capture, as well as persistent and random attacks, by model-based performance analysis validated with extensive simula-tion based on ns3. We also demonstrate the validity of our design by a compara-tive performance analysis with existing multipath routing protocols through ex-tensive simulation.








You are cordially invited to attend Feng Chen's Ph.D. Final Defense
Nov. 30, 2012, 3pm, Room NVC 207



Speaker: Feng Chen
Advisor: Prof. Chang-Tien Lu


Title: Efficient Algorithms for Mining Large Spatio-Temporal Data

Abstract:


Knowledge discovery on spatio-temporal datasets has attracted growing interests. Recent advances on remote sensing technology mean that massive amounts of spatio-temporal data are being collected, and its volume keeps increasing at an ever faster pace. It becomes critical to design efficient algorithms for identifying novel and meaningful patterns from massive spatio-temporal datasets. Different from the other data sources, this data exhibits significant space-time statistical dependence, and the assumption of i.i.d. is no longer valid. The exact modeling of space-time dependence will render the exponential growth of model complexity as the data size increases. This research focuses on the construction of efficient and effective approaches using approximate inference techniques for three main mining tasks, including spatial outlier detection, robust spatio-temporal prediction, and novel applications to real world problems.

Spatial novelty patterns, or spatial outliers, are those observations whose characteristics are markedly different from their spatial neighbors. There are two major branches of spatial outlier detection methodologies, including the global Kriging based and the local Laplacian smoothing based. The former approach requires the exact modeling of spatial dependence, which is time extensive; and the latter approach requires the i.i.d. assumption of the smoothed observations, which is not statistical solid. These two approaches are constrained to numerical data, but in real world applications we are often faced with a variety of non-numerical data types, such as count, binary, nominal, and ordinal. To summarize, the main research challenges are: (1) how to model large data variations caused by outliers; (2) how to effectively and efficiently detect outliers for large numerical spatial datasets; (3) how to generalize numerical detection methods and develop a unified outlier detection framework suitable for large non-numerical datasets; (4) how to achieve accurate spatial prediction even when the training data has been contaminated by outliers; (5) how to deal with spatio-temporal data for the preceding problems.

In this proposed work, dimension-reduced statistical models and approximate inference algorithms will be developed to satisfy the demand of high effectiveness and efficiency for mining large spatial datasets. Proposed distributions of this work include: (1) theoretical development of a generalized local statistics (GLS) model for numerical data; (2) design of two improved forward and backward algorithms for numerical spatial outlier detection; (3) theoretical development of a robust and reduced-rank generalized linear model for non-numerical spatial data; (4) design of a generic approach to non-numerical outlier detection that supports count, binary, ordinal, and nominal data attributes; (5) theoretical development of a robust spatio-temporal random effects model; (6) design of several robust spatio-temporal prediction algorithms. Our proposed statistical models were generalized based on traditional spatial local-statistics, spatial Kriging, and spatio-temporal random effects models. The designed outlier detection and robust prediction algorithms can be mostly executed in near linear time. Two learning algorithms are further designed to address two real world problems, including activity analysis using low sample rate smart meters, and device fingerprinting to enhance wireless security using infinite hidden Markov random field.








You are cordially invited to attend Fenye Bao's Ph.D. Research Defense
Nov. 28, 2012, 4pm, Room NVC T3



Speaker: Fenye Bao
Advisor: Prof. Ing-Ray Chen


Title: Dynamic Trust Management for Mobile Networks and Its Applications

Abstract:


Trust management in mobile networks is challenging due to dynamically changing network environments and the lack of a centralized trusted authority. In this dissertation research, we design and validate a class of dynamic trust man-agement protocols for mobile networks, and demonstrate the utility of dynamic trust management with trust-based applications. Unlike existing work, we con-sider social trust derived from social networks in addition to traditional quality-of-service (QoS) trust derived from communication networks to obtain a composite trust metric as a basis for evaluating trust of nodes in mobile network applica-tions. Untreated in the literature, we design and validate trust composition, ag-gregation, propagation, and formation protocols for dynamic trust management that can learn from past experiences and adapt to changing environment condi-tions to maximize application performance and enhance operation agility. Fur-thermore, we propose, explore and validate the design concept of application-level trust optimization in response to changing conditions to maximize applica-tion performance or best satisfy application requirements. We provide formal proof for the convergence, accuracy, and resiliency properties of our trust man-agement protocols. To achieve the goals of identifying the best trust protocol set-ting and optimizing the use of trust for trust-based applications, we develop a novel model-based analysis methodology with simulation validation for analyz-ing and validating our dynamic trust management protocol design.

The dissertation research provides new understanding of dynamic trust man-agement for mobile wireless networks. We gain insight on the best trust composi-tion and trust formation out of social and QoS trust components, as well as the best trust aggregation and propagation protocols for optimizing application per-formance. We gain insight on how a modeling and analysis tool can be built, al-lowing trust composition, aggregation, propagation, and formation designs to be incorporated, tested and validated. We demonstrate the utility of dynamic trust management protocol for mobile networks including mobile ad-hoc networks, delay tolerant networks, wireless sensor networks, and Internet of things systems with practical applications including misbehaving node detection, trust-based survivability management, trust-based secure routing, and trust-based service composition. Through model-based analysis with simulation validation, we show that our dynamic trust management based protocols outperform existing non-trust-based and trust-based protocols in the presence of malicious, erroneous, partly trusted, uncertain and incomplete information, and are resilient to trust related attacks.








You are cordially invited to attend Robert Mitchell's Ph.D. Research Defense
Nov. 15, 2012, 2pm, Room NVC 401



Speaker: Robert Mitchell
Advisor: Prof. Ing-Ray Chen


Title: Design and Analysis of Intrusion Detection Protocols in Cyber Physical Systems

Abstract:


In this dissertation research we aim to design and validate intrusion detection system (IDS) protocols for a cyber physical system (CPS) comprising sensors, actuators, control units, and physical objects for controlling and protecting physical infrastructures.

The design part includes host IDS, system IDS, and IDS response designs. The validation part includes a novel model based analysis methodology with simulation validation. Our objective is to maximize the CPS reliability or lifetime in the presence of malicious nodes performing attacks which can cause security failures. Our host IDS design results in a lightweight, accurate, autonomous and adaptive protocol that runs on every node in the CPS to detect misbehavior of neighbor nodes based on state-based behavior specifications. Our system IDS design results in a robust and resilient protocol that can cope with malicious, erroneous, partly trusted, uncertain and incomplete information in a CPS. Our IDS response design results in a highly adaptive and dynamic control protocol that can adjust detection strength in response to environment changes in attacker strength and behavior. The end result is an energy-aware and adaptive IDS that can maximize the CPS lifetime in the presence of malicious attacks, as well as malicious, erroneous, partly trusted, uncertain and incomplete information.

We develop a probability model based on stochastic Petri nets to describe the behavior of a CPS incorporating our proposed intrusion detection and response designs, subject to attacks by malicious nodes exhibiting a range of attacker behaviors, including reckless, random, insidious and opportunistic attacker models. We identify optimal intrusion detection settings under which the CPS reliability or lifetime is maximized for each attacker model. Adaptive control for maximizing IDS performance is achieved by dynamically adjusting detection and response strength in response to attacker strength and behavior detected at runtime. We conduct extensive analysis of our designs with four case studies, namely, a mobile group CPS, a medical CPS, a smart grid CPS and an unmanned aircraft CPS. The results show that our adaptive intrusion and response designs operating at optimizing conditions significantly outperform existing anomaly based IDS techniques for CPSs.








You are cordially invited to attend Raimundo F. Dos Santos Jr.'s Ph.D. Preliminary Exam Defense
Sept. 7, 2012, 10am, Room NVC 351



Speaker: Raimundo F. Dos Santos Jr.
Advisor: Prof. Chang-Tien Lu


Title: Data Analysis in Spatial Contexts

Abstract:


With the growing spread of spatial data, exploratory analysis across datasets has gained a considerable amount of attention. Particularly in the fields of Information Retrieval and Data Mining, the integration of data points helps uncover interesting patterns not always visible to the naked eye. Social networks often link entities that share places and activities; marketing tools target users based on behavior and preferences; and medical technology combines symptoms to categorize diseases. Many of the current approaches in this eld of research rely on syntactic heuristics, which are good for comparisons, but less than ideal for inferences. Others apply semantic methods helpful in drawing extended conclusions, but which fail to incorporate syntacts appropriately. This research focuses on spatial data analysis that incorporates both syntactic and semantic methods.

From a functional perspective, any spatial object can be investigated at the entity or at the attribute levels. The former attempts to predict how two entities are alike, correponding to a semantic view; the latter makes no assumptions about the entities as a whole, but rather, observes if any of their attributes are similar, which leans toward a syntactic view. Existing research examines several aspects of entities and their attributes: shared relationships among objects, matches versus mismatches of values, distances among parents and children, and brute-force comparison of characteristics. Most of er from the pitfalls of disparate data, often missing true relationships, failing to deal with inexact vocabularies, ignoring missing values, and poorly handling multiple attributes. In addition, the vast majority does not consider the spatial aspects of the data.

This research combines semantic and syntactic techniques of data analysis in spatial contexts. The proposed solutions are able to identify relationships that are not explicitly written down. They consider both the presence of populated attributes and the cases when values are absent. In addition, they allow comparison of objects erent categorizations and are able to establish similarity for categorical data. Major contributions of this research include (1) a framework that computes a numerical entity similarity, denoted a semantic footprint, composed of spatial, dimensional, and ontological facets; (2) a semantic approach that translates categorical data into a numerical similarity, which permits ranking and ordering; (3) an extensive study of GML as a representative spatial structure of how semantic analysis methods are in uenced by its approaches to storage, querying, and parsing; (4) a design that explores the qualitative aspects of temporal relationships to determine entity similarity; (5) a multi-resolution approach that relates entities categorized under different levels.








You are cordially invited to attend Reghu Anguswamy's Ph.D. Preliminary Exam Defense
May 9, 2012, 2pm, Room NVC 204



Speaker: Reghu Anguswamy
Advisor: Prof. William B. Frakes


Title: A Study of Factors Affecting the Design and Use of Reusable Components

Abstract:


Designing and building components to be reusable is a key area in software reuse research. Practitioners and researchers need to address the problem of how to build reusable components. In this thesis, we will study design principles that can be applied to make components reusable. These design principles are language and domain independent. With an empirical study we will identify the most commonly used reuse design principles. This can be a guideline for designing and building reusable components. Re-engineering a component to be reusable by applying the reuse design principles is a cumbersome task. It is important to understand the overhead involved in making components reusable. In this thesis, we will conduct an empirical study analyzing the overhead in terms of component size, effort required, number of parameters, and productivity. Reusing components in a system involves many challenges. Successful reuse of the components depends on how easily a user can use them in the system. It is important to understand the factors that affect the ease of reuse. Through an empirical study, we will analyze the effect of the size of a component and the reuse design principles used in building the component on the ease of reuse.

We will analyze the human factors that affect the ease of reuse. The human factors studied are the experiences of the user in software programming, software reuse, and programming language. We will also analyze whether component testing makes it easier to reuse or not.








You are cordially invited to attend Feng Chen's Ph.D. Research Defense
May 10, 2012, 2pm, Room NVC 351



Speaker: Feng Chen
Advisor: Prof. Chang-Tien Lu


Title: Efficient Algorithms for Mining Large Spatio-Temporal Data

Abstract:


Nowadays, knowledge discovery on spatio-temporal data sets has attracted growing interests. Recent advances in remote sensing technology mean that massive amounts of spatio-temporal data are now collected, and this volume will only increase. It becomes critical to design ecient algorithms for identifying novel and meaningful patterns from massive spatio-temporal data sets.

Different from other data types, spatio-temporal data usually exhibits statistical dependencies between data objects based on space and time, and the assumption of i.i.d. is no longer valid. The global modeling of dependencies between data objects will render the exponential growth of model complexity with the total number of data objects. An alternative approximation based on local space dependencies has become a promising research direction to balance the tradeoR between model complexity and the size of data sample.

This thesis focuses on the development of local space and geometry based learning techniques for three spatial mining tasks, including spatial outlier detection, robust spatio-temporal prediction, and novel applications to real world problems. For the rst task, existing local based solutions for spatial outlier detection are mostly heuristics driven and lack sucient statistical justications. The proposed contributions of our work include: (1) design of a generalized local statistical (GLS) framework to provide statistical foundations for the family of local based methods; (2) robust estimation and improved outlier detection methods based on the proposed GLS framework; (3) in-depth theoretical evaluations and comprehensive simulations on the comparisons between local and global based detection methods; and (4) extension of the GLS framework to non-numerical spatial data, such as binary, count, and ordinal spatial data.

For the second task, traditional algorithms for predicting spatio-temporal data are mostly designed based on Gaussian process and linear dynamic systems, hence not applicable for the nonlinear dynamic environments popular in many real applications. This deciency can be systematically addressed by increasing the robustness of existing algorithms using heavy tailed distributions, such as the Huber, Laplace, and Student's t distributions. The proposed contributions of our work include: 1) Design of a robust spatio-temporal random eRects model (R-STRE) to capture the spatio- temporal dependence; 2) Formalization of the robust spatio-temporal prediction problem; 3) Design of a general prediction algorithm that can be applied to most existing heavy tailed distributions; 4) Development of optimization techniques for the special Huber and Laplace distributions; and 5) Comprehensive experiments to validate the new algorithm's robustness and efficiency.

For the third task, we present novel applications of spatio-temporal mining to two real world probems, including activity analysis based on low sample rate smart meters, and device ngerprinting to enhance wireless security using innite hidden Markov random field.








You are cordially invited to attend Yinan Li's Ph.D. Final Defense
April. 30, 2012, 11am, Room NVC 351



Speaker: Yinan Li
Advisor: Prof. Ing-Ray Chen


Title: Integrated Mobility and Service Management for Network Cost Minimization in Wireless Mesh Networks

Abstract:


In this dissertation research, we design and analyze integrated mobility and service manage- ment for network cost minimization in Wireless Mesh Networks (WMNs). We first investigate the problem of mobility management in WMNs for which we propose two efficient per-user mobility management schemes based on pointer forwarding, and then a third one that integrates routing- based location update and pointer forwarding for further performance improvement.

We further study integrated mobility and service management for which we propose protocols that support efficient mobile data access services with cache consistency management, and mobile multicast services. We also investigate reliable and secure integrated mobility and service man- agement in WMNs, and apply the idea to the design of a protocol for secure and reliable mobile multicast. The most salient feature of our protocols is that they are optimal on a per-user basis (or on a per-group basis for mobile multicast), that is, the overall network communication cost incurred is minimized for each individual user (or group). Per-user based optimization is critical because mobile users normally have vastly different mobility and service characteristics. Thus, the overall cost saving due to per-user based optimization is cumulatively significant with an increasing mobile user population.

To evaluate the performance of our proposed protocols, we develop mathematical models and computational procedures used to compute the network communication cost incurred and build simulation systems for validating the results obtained from analytical modeling. We identify optimal design settings under which the network cost is minimized for our mobility and service management protocols in WMNs. Intensive comparative performance studies are carried out to compare our protocols with existing work in the literature. The results show that our protocols significantly outperform existing protocols under identical environmental and operational settings.

We extend the design notion of integrated mobility and service management for cost minimiza- tion to MANETs and propose a scalable dual-region mobility management scheme for location- based routing. The basic design concept is to use local regions to complement home regions and have mobile nodes in the home region of a mobile node serve as location servers for that node. We develop a mathematical model to derive the optimal home region size and local region size under which overall network cost incurred is minimized. Through a comparative performance study, we show that dual-region mobility management outperforms existing mobility management schemes based on static home regions.







You are cordially invited to attend Iman Saleh Moustafa's Ph.D. Final Defense
Feb. 10, 2012, 2:30pm, Room NVC 204



Speaker: Iman Saleh Moustafa
Advisor: Prof. Gregory W. Kulczycki


Title: Formal Specification and Verification of Data-Centric Web Services

Abstract:


In this thesis, we develop and evaluate a formal model and contracting framework for data-centric Web services. The central component of our framework is a formal specification of a common Create- Read-Update-Delete (CRUD) data store. We show how this model can be used in the formal specification and verification of both basic and transactional Web service compositions. We demonstrate through both formal proofs and empirical evaluations that our proposed framework significantly decreases ambiguity about a service, enhances its reuse, and facilitates detection of errors in servicebased implementations. Web Services are reusable software components that make use of standardized interfaces to enable loosely-coupled business-to-business and customer-to-business interactions over the Web. In such environments, service consumers depend heavily on the service interface specification to discover, invoke, and synthesize services over the Web. Data-centric Web services are services whose behavior is determined by their interactions with a repository of stored data. A major challenge in this domain is interpreting the data that must be marshaled between consumer and producer systems. While the Web Services Description Language (WSDL) is currently the de facto standard for Web services, it only specifies a service operation in terms of its syntactical inputs and outputs; it does not provide a means for specifying the underlying data model, nor does it specify how a service invocation affects the data. The lack of data specification potentially leads to erroneous use of the service by a consumer. In this work, we propose a formal contract for data-centric Web services. The goal is to formally and unambiguously specify the service behavior in terms of its underlying data model and data interactions. We address the specification of a single service, a flow of services interacting with a single data store, and also the specification of distributed transactions involving multiple Web services interacting with different autonomous data stores. We use the proposed formal contract to decrease ambiguity about a service behavior, to fully verify a composition of services, and to guarantee correctness and data integrity properties within a transactional composition of services.







You are cordially invited to attend Fenye Bao's Ph.D. Preliminary Exam Defense
Nov. 21, 2011, 3pm, Room NVC 325



Speaker: Fenye Bao
Advisor: Dr. Ing-Ray Chen


Title: Dynamic Trust Management for Mobile Networks and Its Applications

Abstract:


Trust management in mobile networks is challenging due to dynamically changing network environments and the lack of a centralized trusted authority. In this dissertation research, we design and validate a dynamic trust management protocol that can provide a subjective yet accurate assessment of trust of mobile nodes by adapting to changing environment conditions, and demonstrate the utility of dynamic trust management in trust-based applications. Unlike existing work, we consider social trust derived from social networks in addition to traditional quality-of-service (QoS) trust derived from communication networks to obtain a composite trust metric as a basis for evaluating trust of nodes in mobile network applications. Untreated in the literature, we design and validate trust composition, aggregation, propagation, formation and revocation protocols for dynamic trust management that can learn from past experiences and adapt to changing environment conditions to maximize application performance and enhance operation agility. Furthermore, we propose, explore and validate the design concept of application-level trust optimization in response to changing conditions to maximize application performance or best satisfy application requirements. To achieve the goals of identifying the best trust protocol setting and optimizing the use of trust for trust-based applications, we develop a novel model-based analysis methodology with simulation validation for analyzing and validating our dynamic trust management protocol design.

The dissertation research provides new understanding of dynamic trust management for mobile wireless networks. We gain insight on the best trust composition and trust formation out of social and QoS trust components, as well as the best trust aggregation and propagation protocols for optimizing application performance, when given applications characteristics and trustee properties as input. We gain insight on how a modeling and analysis tool can be built, allowing trust composition, aggregation, propagation, formation and revocation designs to be incorporated, tested and validated. We demonstrate the utility of dynamic trust management protocol for mobile networks including mobile ad-hoc networks, delay tolerant networks, and wireless sensor networks with practical applications including misbehaving node detection, trust-based survivability management, and trust-based secure routing. Through model-based analysis with simulation validation, we show that our dynamic trust management based protocols outperform existing non-trust-based and trust-based protocols in the presence of malicious, erroneous, partly trusted, uncertain and incomplete information, and are resilient to trust related attacks.







You are cordially invited to attend Xutong Liu's Ph.D. Preliminary Exam Defense
Nov. 16, 2011, 2pm, Room NVC 351



Speaker: Xutong Liu
Advisor: Prof. Chang-Tien Lu


Title: Outlier Detection Techniques in Spatial Dataset

Abstract:


With the ever-increasing volume of spatial data, identifying hidden but potentially interesting patterns has attracted considerable attentions, particularly from the areas of data mining experts and geographers. Spatial outlier analysis, which aims at detecting abnormal objects in spatial contexts, becomes one of the important spatial data mining branches. The identication of spatial outliers can help extract important knowledge in many applications, including meteorological data analysis, tra c control, satellite image analysis, geological data mining and hotspot identication. Although most spatial outlier detection approaches have been proposed, there exist the well-known masking and swamping problems, and most of them can't still satisfy certain requirements aroused recently. This report focuses on the development of spatial outlier detection techniques for three aspects, including spatial numerical outlier detection, spatial categorical outlier detection and identication of the number of spatial numerical outliers.

Firstly, the benets of random walk techniques are investigated on computing the relevance among spatial numerical objects. Within this framework, two random walk based approaches, RW- BP(Random Walk on Bipartite Graph) and RW-EC(Random Walk on Exhaustive Combination), erent weighted graphs, a BP(Bipartite Graph) and an EC(Exhaustive Combination), are modeled based on the spatial and/or non-spatial attributes of the spatial objects. Then, random walk techniques are utilized on the graphs to compute the relevance scores between the spatial objects. Using the analysis results, the outlier scores are com- puted for each object and the objects with higher outlierness are recognized as outliers.

When encountering categorical dataset, some introduce spatial numerical outlier detection methods by directly mapping the categorical attributes to continuous ones. However, such solutions fail to capture the specic properties of spatial categorical data, which is prone to incur the masking and swamping issues. Therefore, the second part focuses on modeling the spatial dependencies between spatial categorical observations and propose several pair Correlation Function based methods to detect spatial categorical outliers in both single and multiple attribute domains. First, a new metric named Pair Correlation Ratio(PCR) is dened and estimated for each pair of categorical combinations based on their co-occurrence frequency erent spatial distances. The outlier score for each object is computed as the average PCRs between itself and its spatial neighbors. The observations with the lowest PCRs are diagnosed as potential SCOs.

Estimating an appropriate number of outliers for a spatial data set is always one of the critical issues for outlier analysis. Finally, this work proposes an entropy-based method to address this problem in spatial numerical domain. According the entropy theory, that is, the data set with more outiers has a higher entropy value than that with less outliers, we expect that, by incrementally removing outliers, the entropy value will decrease sharply, and reach a stable state when all the outliers have been removed. Experimental evaluation on a real dataset by integrating proposed method with POD approach to demonstrate its effectiveness.







You are cordially invited to attend Robert Mitchell's Ph.D. Preliminary Exam Defense
Nov. 10, 2011, 10am, Room NVC 351



Speaker: Robert Mitchell
Advisor: Dr. Ing-Ray Chen


Title: Design and Analysis of Intrusion Detection Protocols in Cyber Physical Systems

Abstract:


In this dissertation research we aim to design and validate intrusion detection system (IDS) protocols for a cyber physical system (CPS) comprising sensors, actuators, control units, and physical objects for controlling and protecting physical infrastructures.

The design part includes host IDS, system IDS, and IDS response designs. The validation part includes a novel model based analysis methodology with simulation validation. Our objective is to maximize the CPS reliability or lifetime in the presence of malicious nodes performing attacks which can cause security failures. We take an incremental design approach, starting with intrusion detection and response protocol design for single enclave CPSs and then progressing to multiple enclave federated CPSs. Our host IDS design results in a lightweight, accurate, autonomous and adaptive protocol that runs on every node in the CPS to detect misbehavior of neighbor nodes based on state-based behavior specifications. Our system IDS design results in a robust and resilient protocol that can cope with malicious, erroneous, partly trusted, uncertain and incomplete information in a CPS. Our IDS response design results in a highly adaptive and dynamic control protocol that can adjust detection strength in response to environment changes in attacker strength and behavior. The end result is an energy-aware and adaptive IDS that can maximize the CPS lifetime in the presence of malicious attacks, as well as malicious, erroneous, partly trusted, uncertain and incomplete information.

We develop a probability model based on stochastic Petri nets to describe the behavior of the CPS incorporating our proposed intrusion detection and response designs, subject to attacks by malicious nodes exhibiting a range of attacker behaviors, including persistent, random, insidious and oracle attacker models. We identify optimal intrusion detection settings under which the CPS reliability or lifetime is maximized for each attacker model. Adaptive control for maximizing IDS performance is achieved by dynamically adjusting detection and response strength in response to attacker strength and behavior detected at runtime. We conduct extensive analysis of our designs. The results show that our adaptive intrusion and response designs operating at optimizing conditions significantly outperform existing anomaly based IDS techniques for CPSs.







You are cordially invited to attend Iman Saleh Moustafa's Ph.D. Research Defense
Oct. 24, 2011, 11am, Room NVC 204



Speaker: Iman Saleh Moustafa
Advisor: Prof. Gregory W. Kulczycki


Title: Formal Specification and Verification of Data-Centric Web Services

Abstract:


In this thesis, we develop and evaluate a formal model and contracting framework for data-centric Web services. The central component of our framework is a formal specification of a common Create- Read-Update-Delete (CRUD) data store. We show how this model can be used in the formal specification and verification of both basic and transactional Web service compositions. We demonstrate through both formal proofs and empirical evaluations that our proposed framework significantly decreases ambiguity about a service, enhances its reuse, and facilitates detection of errors in servicebased implementations. Web Services are reusable software components that make use of standardized interfaces to enable loosely-coupled business-to-business and customer-to-business interactions over the Web. In such environments, service consumers depend heavily on the service interface specification to discover, invoke, and synthesize services over the Web. Data-centric Web services are services whose behavior is determined by their interactions with a repository of stored data. A major challenge in this domain is interpreting the data that must be marshaled between consumer and producer systems. While the Web Services Description Language (WSDL) is currently the de facto standard for Web services, it only specifies a service operation in terms of its syntactical inputs and outputs; it does not provide a means for specifying the underlying data model, nor does it specify how a service invocation affects the data. The lack of data specification potentially leads to erroneous use of the service by a consumer. In this work, we propose a formal contract for data-centric Web services. The goal is to formally and unambiguously specify the service behavior in terms of its underlying data model and data interactions. We address the specification of a single service, a flow of services interacting with a single data store, and also the specification of distributed transactions involving multiple Web services interacting with different autonomous data stores. We use the proposed formal contract to decrease ambiguity about a service behavior, to fully verify a composition of services, and to guarantee correctness and data integrity properties within a transactional composition of services.







You are cordially invited to attend Daniel Wayne Smith's MS Thesis Defense
Dec. 7, 2010



Speaker: Daniel Wayne Smith
Advisor: Prof. Gregory Kulczycki


Title: Preserving Unique References in Java Lists

Abstract:


The Java collection framework introduces aliasing when objects are added to and accessed from collections. This thesis describes a list component implemented in Java that preserves unique references of objects in the list, thereby avoiding undesired aliasing. We compared the running time of our list with three other lists from Java collections (Java collection framework, Google, and Functional Java) in five different applications. We found that the performance of our list was usually slightly slower than the performance of the Java list, but often much faster than the Google and Functional Java lists. We also compared the reasoning complexity of our list with Java’s list by creating tracing tables for a method from a towers-of-Hanoi application and comparing the number of tokens in the table using our list with the number of tokens in the table using the Java list. We found that the number of tokens in the tracing table using the Java list was much higher than the number of tokens in the table using our list. We argue that this result will occur in any table for applications that use mutable list objects.







You are cordially invited to attend Yinan Li's Ph.D. Preliminary Exam Defense
October 27, 2010, 10am, Room 320



Speaker: Yinan Li
Advisor: Prof. Ing-Ray Chen


Title: Integrated Mobility and Service Management for Cost Minimization in Wireless Mesh Networks

Abstract:


Wireless mesh networks (WMNs) are emerging in recent years as a key driving technology for next-generation wireless networks, and are widely regarded as a cost-effective solution for providing last-mile broadband Internet access on a community basis. As WMNs have networking characteristics drastically different from mobile ad hoc networks (MANETs) and wireless sensor networks (WSNs), existing network protocols designed for MANETs and WSNs cannot be ported to WMNs. To realize the full potential of WMNs so as to maximize the throughput or to minimize the overall communication cost of various network services in WMNs, we need protocols that are particularly designed and developed for WMNs, taking into consideration their unique networking characteristics, including a semi-static mesh routing backbone, one or more gateways connecting the WMN to the Internet, and the existence of relatively powerful mesh routers on the mesh routing backbone executing routing and location management functions. As the MANET technology is new in recent years, there is not much research work on mobility and service management, which is considered key to its success. This dissertation research aims at filling the gap.

In this dissertation research, we design and analyze several mobility management protocols with the objective of minimizing the overall network cost, and develop the notion of integrated mobility and service management for key applications in WMNs, including mobile data access and mobile multicast. More specifically, we first investigate the problem of mobility management in WMNs for which we propose two efficient per-user based schemes that minimize the overall network communication cost incurred collectively by mobility management and packet delivery. Mobility management is critical for the proper operation of a WMN and is the basis for uninterrupted network services because it maintains necessary information for a service to be delivered to a mesh client that changes its location frequently. We further study per-user based integrated mobility and service management for which we propose protocols that support efficient mobile data access services with cache consistency management, and mobile multicast services. The most salient feature of our protocols is that they are optimal on a per-user basis (or on a per-group basis for mobile multicast), that is, the overall network communication cost incurred is minimized for each individual user (or group). Per-user based optimization is critical because mobile users normally have vastly different mobility and service characteristics. Thus, the overall cost saving due to per-user based optimization is cumulatively significant with an increasing mobile user population.

To evaluate the performance of our proposed protocols, we develop mathematical models based on stochastic Petri net techniques and build simulation systems for validating the results obtained from analytical modeling. We identify optimal design settings under which the network cost is minimized for our mobility and service management protocols in WMNs. Intensive comparative performance studies are carried out to compare our protocols with existing work in the literature. The results show that our protocols significantly outperform existing protocols under identical environmental and operational settings. We demonstrate that our design notion of integrated mobility and service management greatly contributes to network cost minimization for WMNs to provide mobile data access and multicast services to mobile users with vastly diverse mobility and service characteristics.






You are cordially invited to attend Feng Chen's Ph.D. Preliminary Exam Defense
October 5, 2010



Speaker: Feng Chen
Advisor: Prof. Chang-Tien Lu


Title: On Local Based Algorithms for Mining Spatial Data

Abstract:


Nowadays, knowledge discovery on spatial data has attracted growing interests. Different from other data types, spatial data usually exhibit statistical dependencies between data objects based on space, and the assumption of i.i.d. is no longer valid. The global modeling of dependencies between data objects will render the exponential growth of model complexity with the total number of data objects. An alternative approximation based on local space dependencies has become a promising research direction to balance the tradeoff between model complexity and the size of data sample.

This thesis focuses on the development of local space and geometry based learning techniques for three spatial mining tasks, including spatial outlier detection, anomalous cluster detection, and locally linear classification. For the first task, existing local based solutions for spatial outlier detection are mostly heuristics driven and lack sufficient statistical justifications. The proposed contributions of our work include: (1) design of a generalized local statistical (GLS) framework to provide statistical foundations for the family of local based methods; (2) robust estimation and improved outlier detection methods based on the proposed GLS framework; (3) in-depth theoretical evaluations and comprehensive simulations on the comparisons between local and global based detection methods; and (4) extension of the GLS framework to multivariate spatial data.

For the second task, traditional solutions for anomalous cluster detection, named spatial scan methods, were designed based on the restrictive assumption that the data follows a mixture of two Gaussians, and did not consider the situations where the data may exhibit complex global trends, significant spatial autocorrelations, and multiple anomalous clusters. The proposed contributions of our work for this task include: (1) design of a local spatial scan (LSS) statistics framework to improve the efficiency and effectiveness of existing spatial scan methods; (2) In-depth evaluations about the statistical properties of the local spatial scan framework; and (3) development of efficient algorithms to search for anomalous clusters based on LSS. For the third and final task, local based linear classifications have been popular techniques used in other domains but rarely in spatial fields. The proposed contributions for this work include: (1) theoretical evaluations on major properties of existing locally linear classification methods; and (2) generalization of this local learning approach to spatial data.






You are cordially invited to attend Arnold P. Boedihardjo's Ph.D. Final Defense
August 10, 2010, 2pm, Room 351



Speaker: Arnold P. Boedihardjo
Advisor: Prof. Chang-Tien Lu


Title: Efficient Algorithms for Mining Data Streams

Abstract:


Data streams are ordered sets of values that are fast, continuous, mutable, and potentially unbounded. Examples of data streams are the pervasive time series, which span domains such as finance, medicine, and transportation. Mining data streams require approaches that are efficient, adaptive, and scalable and for several stream mining tasks, knowledge of the datas probability density function (PDF) is essential to deriving usable results. Providing an accurate model for the PDF benefits a variety of stream mining applications and its successful development can have far-reaching impact to the general discipline of stream analysis. Therefore, this research focuses on the construction of efficient and effective approaches for estimating the PDF of data streams.

In this work, kernel density estimators (KDEs) will be developed that satisfy the stringent computational stipulations of data streams, model unknown and dynamic distributions, and enhance the estimation quality of complex structures. Contributions of this work include: (1) theoretical development of the local region based KDE; (2) construction of a local region based estimation algorithm for data streams; (3) design of a generalized local region approach that can be applied any existing KDE; and (4) application extension of the local region based KDE to multi-scale outlier detection. Theoretical development includes the formulation of the local region concept to effectively approximate the accurate but computationally intensive adaptive KDE. The work also derives and analyzes key theoretical properties of the local region based approach which include (amongst others) its expected performance, an alternative local region construction criterion, and its viability as a general framework to enhance the accuracy of existing KDEs. Algorithmic design includes the development of a specific estimation technique that reduces the time/space complexities of the adaptive KDE. A technique for optimizing multiple density queries is also proposed to reduce the computations required to perform data stream monitoring tasks such outlier detection and visualization. The local region concept is extended to an efficient algorithmic framework which can be applied to existing stream-based KDEs to enhance estimation quality. As an application extension, a multi-scale outlier detection framework is designed which can effectively detect anomalies within a hierarchy of sliding windows.







You are cordially invited to attend Chaitanya Nemallapudi's MS Thesis Defense
August 2, 2010, 2pm, Room 325



Speaker: Chaitanya Nemallapudi
Advisor: Prof. Bill Frakes and Dr. Greg Kulczycki


Title: Comparison of Domain Vocabularies Across Domain Experts

Abstract:


Vocabularies created by various domain experts working on the same domain (conflation algorithms), were compared to check for consistency. Sources used by each domain expert were also compared. The measure of comparison is termed as "Overlap". The results of the tests show that the mean overlap score of the vocabularies and the mean overlap score of source documents were significantly greater than zero. However, the overlap scores of the vocabularies and the source documents were not significantly different between the domain experts. The relationship between the overlap of vocabularies and the source documents was also evaluated. The results show a weak correlation between them. Also, the variability of the vocabularies generated automatically to the variability of those produced manually by domain experts was evaluated. The results suggested that the vocabularies are significantly different from each other.







You are cordially invited to attend Donald W. McCormick II 's MS Thesis Defense
Apr. 28, 2010, 2pm, Room 325



Speaker: Donald W. McCormick II
Advisor: Prof. Bill Frakes


Title: A Sufficient Set of Mutation Operators for Structured Query Language (SQL)

Abstract:


Mutation testing involves systematically inserting artificial faults in program code or SQL in order to mimic real faults. The fault-detection capability, or mutation adequacy score (AM), of test suites on these artificial faults, or mutants, compares to the fault-detection capability of test suites on real faults. A test suite's AM is the ratio of mutants detected to the total number of non-equivalent mutants. Mutation analysis involves selecting a sufficient set of mutation operators that minimize the number of mutants necessary to accurately predict a test suite's AM. A set of mutation operators has previously been identified for SQL clause, operator replacement, null handling, and identifier replacement categories. The focus of this paper is on determining whether the current set of SQL mutation operators is sufficient. Experiments are conducted that reveal poor actors - SQL mutation operators that consistently contribute low scores to the overall test suite AM. Additional SQL mutation operators are introduced for each poor actor category that improve the overall test suite AM.







You are cordially invited to attend Arnold P. Boedihardjo's Ph.D. Research Defense
Jan. 27, 2010, 2pm, Room 313



Speaker: Arnold P. Boedihardjo
Advisor: Prof. Chang-Tien Lu


Title: Efficient Algorithms for Mining Data Streams

Abstract:


Advances in hardware and software technologies have caused a surge in the growth and prevalence of data streams. Examples of data streams are the pervasive time series, which span domains such as finance, medicine, and transportation. Data streams are ordered sets of values that are fast, continuous, mutable, and potentially unbounded. As implied by their properties, processing data streams require approaches that are efficient, adaptive, and scalable. Any applied mining tool must therefore heed this same set of processing requirements. For several stream mining problems, knowledge of the underlying data distribution is the primal requirement for deriving usable results. Because the data distribution is completely described by the probability density function, providing a model for the density function benefits a variety of stream mining applications and its successful development can have far-reaching impact to the general discipline of stream analysis. Therefore, this research focuses on the construction of efficient and effective approaches for estimating the probability density function of data streams.

The problems of developing a viable probability density estimator are rooted at the data streams large size, fast arrival rate, unknown form, and evolving distribution. Due to its dynamic nature, traditional parametric modeling methods are not suited to describe the evolving structure of data streams. Hence, nonparametric techniques are considered to effectively characterize their underlying probability density function. The nonparametric kernel density estimator is chosen as the focus of this work. The kernel density estimator possesses attractive characteristics that include its ability to generalize other estimators, well-established theoretical properties, asymptotic consistency, and inheritance of the kernel functions continuity and differentiability. In its native form, the estimators potentially high computational and memory requirements can render their implementations infeasible under the data streams high arrival rate and massive size. Existing stream-based kernel estimators offer promise in addressing some of these deficiencies, however, their solutions are attained at the cost of significant reduction in estimation quality. When faced with a complex distribution, existing techniques can produce oversmoothed estimates and miss critical features which result in significant estimation errors.

In this proposed work, kernel density estimators will be developed that satisfy the stringent computational stipulations of data streams, model unknown and dynamic distributions, and enhance the estimation quality of complex structures. Proposed contributions of this work include: (1) theoretical development of the local region based kernel density estimator; (2) construction of a local region based estimation algorithm for data streams; (3) design of a generalized local region approach that can be applied an existing estimator; and (4) application extension of the local region based estimator to the transportation domain. Theoretical development includes the formulation of the local region concept to effectively approximate the accurate but computationally intensive adaptive kernel density estimator. The work also derives and analyzes key theoretical properties of the local region based approach which include (amongst others) its expected performance, an alternative local region construction criterion, and its viability as a general framework to enhance the accuracy of existing solutions. Algorithmic design will include the development of a specific nonparametric estimation technique that reduces the time/space complexities of the adaptive kernel density estimator under a sliding window. The local region concept is extended to an efficient algorithmic framework which can be applied to existing stream-based estimators to enhance estimation quality. As an application extension, an incident detection system is designed and implemented which can efficiently detect non-recurrent congestions and adapts to changing traffic conditions. A technique for optimizing multiple density queries is also proposed to reduce the computations required to perform data stream monitoring tasks such as visualization and outlier detection.







You are cordially invited to attend Iman Saleh Moustafa 's PhD Preliminary Exam Defense
Dec. 4, 2009, 10am, Room 207



Speaker: Iman Saleh Moustafa
Advisor: Dr. Gregory W. Kulczycki


Title: Specification and Verification of Data-Centric Web Services

Abstract:


Web Services are reusable software components that make use of standardized interfaces to enable loosely-coupled business-to-business and customer-to-business interactions over the Web. In such environments, service consumers depend heavily on the service interface specification to discover, invoke, and synthesize services over the Web. Data-centric Web services are services whose behavior is determined by their interactions with a repository of stored data. A major challenge in this domain is interpreting the data that must be marshaled between consumer and producer systems. While the Web Services Description Language (WSDL) is currently the de facto standard for Web services, it only specifies a service operation in terms of its syntactical inputs and outputs; it does not provide a means for specifying the underlying data model, nor does it specify how a service invocation affects the data. The lack of data specification potentially leads to erroneous use of the service by a consumer. In this work, we propose a formal contract for data-centric Web services. The goal is to formally and unambiguously specify the service behavior in terms of its underlying data model and data interactions. We address the specification of a single service, a flow of services interacting with a single data store, and also the specification of distributed transactions involving multiple Web services interacting with different autonomous data stores. We use the proposed formal contract to decrease ambiguity about a service behavior, to fully verify a composition of services, and to guarantee correctness and data integrity properties within a transactional composition of services.







You are cordially invited to attend Jing Dai 's Ph.D. Final Defense
Sept. 4, 10am, 2009, Room 351



Speaker: Jing Dai
Advisor: Prof. Chang-Tien Lu


Title: Efficient Concurrent Operations in Spatial Databases

Abstract:


As demanded by applications such as GIS, CAD, ecology analysis, and space research, efficient spatial data access methods have attracted much research. Especially, moving object management and continuous spatial queries are becoming highlighted in the spatial database area. However, most of the existing spatial query processing approaches were designed for single-user environments, which may not ensure correctness and data consistency in multiple-user environments. This research focuses on designing efficient concurrent operations on spatial datasets.

Current multidimensional data access methods can be categorized into two types: 1) pure multidimensional indexing structure such as the R-tree family and grid file; 2) linear spatial access methods, represented by the Space-Filling Curve (SFC) combined with B-trees. Concurrency control protocols have been designed for some pure multidimensional indexing structures, but none of them is suitable for variants of R-trees with object clipping, which are efficient in searching. On the other hand, there is no concurrency control protocol designed for linear spatial indexing structures, where the one-dimensional concurrency control protocols cannot be directly applied. Furthermore, the recently designed query processing approaches for moving objects have not been protected by any efficient concurrency control protocols.

In the proposed research, sound solutions for efficient concurrent access frameworks on both types of spatial indexing structures are provided, as well as the continuous query processing on moving objects, for multiple-user environments. These concurrent access frameworks can satisfy the concurrency control requirements, meanwhile provide outstanding performance for concurrent queries. Major contributions of this research include: (1) a new efficient spatial indexing approach with object clipping technique, ZR+-tree, that outperforms R-tree and R+-tree on searching; (2) a concurrency control protocol, GLIP, to provide high throughput and phantom update protection on spatial indexing with object clipping; (3) efficient concurrent operations for indices based on linear spatial access methods, which form up the CLAM protocol; (4) efficient concurrent continuous query processing on moving objects for both R-tree-based and linear spatial indexing frameworks; (5) a generic access framework, Disposable Index, for optimal location update and parallel search.







You are cordially invited to attend Weiping He 's Ph.D. Final Defense
March 20, 10am, 2009, Room 322



Speaker: Weiping He
Advisor: Prof. Ing-Ray Chen


Title: Mobility and Service Management for Future All-IP based Wireless Networks

Abstract:


Mobility management addresses the issues of how to track and locate a mobile node (MN) efficiently. Service management addresses the issues of how to efficiently deliver services to MNs. This dissertation aims to design and analyze integrated mobility and service management schemes for future all-IP based wireless systems. We propose and analyze per-user regional registration schemes extending from Mobile IP Regional Registration and Hierarchical Mobile IPv6 for integrated mobility and service management with the goal to minimize the network signaling and packet delivery cost in future all-IP based wireless networks. We show that, when given a set of parameters characterizing the operational and workload conditions of a MN, there exists an optimal regional area size for the MN such that the network communication cost is minimized for serving mobility and service management operations of the MN.

If access routers in future all-IP based wireless networks are restricted to perform network layer functions only, we investigate the design of intelligent routers, called dynamic mobility anchor points (DMAPs), to implement per-user regional management in IP wireless networks. These DMAPs are access routers (ARs) chosen by individual MNs to act as regional routers to reduce the signaling overhead for intra-regional movements. The DMAP domain size, i.e., the number of subnets covered by a DMAP, is based on a MN's mobility and service characteristics. A MN optimally determines when and where to launch a DMAP to minimize the network cost in serving the user's mobility and service management operations. We show that there exists an optimal DMAP domain size for each individual MN. We also demonstrate that the DMAP design can easily support failure recovery because of the flexibility of allowing a MN to choose any AR to be the DMAP for mobility and service management.

If access routers are powerful and flexible in future all-IP based networks to perform network-layer and application-layer functions, we propose the use of per-user proxies that can run on access routers. The user proxies can carry service context information such as cached data items and Web processing objects, and perform context-aware functions such as content adaptation for services engaged by the MN to help application executions. We investigate a proxy-based integrated mobility and service management architecture (IMSA) under which a client-side proxy is created on a per-user basis to serve as a gateway between a MN and all services engaged by the MN. Leveraging Mobile IP with route optimization, the proxy runs on an access router and cooperates with the home agent and foreign agent of the MN to maintain the location information of the MN to facilitate data delivery by services engaged by the MN. Further, the proxy optimally determines when to move with the MN so as to minimize the network cost associated with the user's mobility and service management operations.

Finally we investigate a proxy-based integrated cache consistency and mobility management scheme called PICMM to support client-server query-based mobile applications, To improve query performance, the MN stores frequently used data in its cache. The MN's proxy receives invalidation reports or updated data objects from application servers, i.e., corresponding nodes (CNs) for cached data objects stored in the MN. If the MN is connected, the proxy will forward invalidation reports or fresh data objects to the MN. If the MN is disconnected, the proxy will store the invalidation reports or fresh data objects, and, once the MN is reconnected, the proxy will forward the latest cache invalidation report or data objects to the MN. We show that there is an optimal ``service area'' under which the overall cost including query processing cost and mobility management cost is minimized. To further reduce network traffic, we develop a threshold-based hybrid cache consistency management policy such that whenever a data object is updated at the server, the server sends an invalidation report to the MN through the proxy to invalidate the cached data object only if the size of the data object exceeds the given threshold. Otherwise, the server sends a fresh copy of the data object through the proxy to the MN. We identify the best ``threshold'' value that would minimize the overall network traffic incurred due to mobility management, cache consistency management, and query processing, when given a set of parameter values characterizing the operational and workload conditions of the MIP network.

We develop mathematical models to analyze performance characteristics of DMAP, IMSA and PICMM developed in the dissertation research and demonstrate that they outperform existing schemes that do not consider integrated mobility and service management or that use static regional routers to serve all MNs in the system. The analytical results obtained are validated through extensive simulation. We discuss design principles and report scientific discoveries resulting from applying these design principles to the development of DMAP, IMSA and PICMM for integrated mobility and service management for future all-IP wireless networks.







You are cordially invited to attend Arnold P. Boedihardjo's PhD Preliminary Exam Defense
Jan. 28, 10am, 2009. Room 313



Speaker: Arnold P. Boedihardjo
Advisor: Prof. Chang-Tien Lu


Title: Efficient Algorithms for Mining Data Streams

Abstract:


Advances in hardware and software technologies have caused a surge in the growth and prevalence of data streams. Examples of data streams are the pervasive time series, which span domains such as finance, medicine, and transportation. Data streams are ordered sets of values that are fast, continuous, mutable, and potentially unbounded. As implied by their properties, processing data streams require approaches that are efficient, adaptive, and scalable. Any applied mining tool must therefore heed this same set of processing requirements. For several stream mining problems, knowledge of the underlying data distribution is the primal requirement for deriving usable results. Because the data distribution is completely described by the probability density function, providing a model of the function benefits a variety of stream mining applications and its successful development can have far-reaching impact to the general discipline of stream analysis. Therefore, this research focuses on the construction of efficient and effective approaches for estimating the probability density function of data streams.

The problems of developing a viable probability density estimator are rooted at the data streams large size, fast arrival rate, unknown form, and evolving distribution. Due to its dynamic nature, traditional parametric modeling methods are not suited to describe the inconstant structure of data streams. Hence, nonparametric techniques are considered to effectively characterize their underlying probability density function. The nonparametric kernel density estimator is chosen as the focus of this work. The kernel density estimator possesses attractive characteristics that include its ability to generalize other estimators, well-studied theoretical properties, asymptotic consistency, and inheritance of the kernel functions continuity and differentiability. In its native form, the estimators potentially high computational and memory requirements can render their implementations infeasible under the data streams high arrival rate and massive size. Existing stream-based kernel estimators offer promise in addressing some of the deficiencies of traditional methods, however, these solutions are attained at the cost of significant reduction in estimation quality. When faced with a complex distribution, existing techniques can produce oversmoothed estimates and miss critical features that result in significant estimation errors.

In this proposed work, kernel density estimators will be developed that satisfy the stringent computational stipulations of data streams, model unknown and dynamic distributions, and enhance the estimation quality of complex structures. Proposed contributions of this work include: (1) theoretical development of the local region based kernel density estimator; (2) algorithmic design of the estimator for data streams; and (3) application extensions to the transportation domain. Theoretical development includes the formulation of the local region concept to effectively approximate the accurate but computationally intensive adaptive kernel density estimator. The work also derives and analyzes key theoretical properties of the local region based approach which include (amongst others) its expected performance, an alternative local region construction criterion, and its viability as a general framework to enhance the accuracy of existing solutions. Algorithmic design will include the development of a specific nonparametric estimation technique that reduces the time/space complexities of the adaptive kernel density estimator under a sliding window. The local region concept is extended to an efficient algorithmic framework by which several existing stream-based estimators can be integrated to enhance overall estimation quality. A technique for optimizing multiple density queries is also proposed to reduce the operations necessary for tasks such as concept drift detection. As an application extension, the local region density estimator is proposed to address the problem of predicting travel times under the conditions of a traffic incident. Lastly, an incident detection system is designed and implemented which can efficiently detect non-recurrent congestions and adapts to changing traffic conditions.







You are cordially invited to attend Jason’s MS Thesis Defense
Dec. 22, 1pm, 2008, Room 325



Speaker: Jason Tilley
Advisor: Prof. Bill Frakes


Title: A Comparison of Statistical Filtering Methods for Automatic Term Extraction for the Purpose of Domain Engineering

Abstract:


Fourteen word frequency metrics were tested to evaluate their effectiveness in identifying vocabulary in a domain. Fifteen domain engineering projects were examined to measure how closely the vocabularies selected by the fourteen word frequency metrics were to the vocabularies produced by domain engineers. Six filtering mechanisms were also evaluated to measure their impact on selecting proper vocabulary terms. The results of the experiment show that stemming and stop word removal do improve overlap scores and that term frequency is a valuable contributor to overlap. Variations on term frequency are not always significant improvers of overlap.







You are cordially invited to attend Arun’s MS Thesis Defense
Dec. 3, 11pm, 2008, Room 324



Speaker: Arun Sudhir
Advisor: Prof. Greg Kulczycki


Title: Tree Component alternatives to the Composite Design Pattern

Abstract:


The Composite design pattern is commonly employed in object-oriented languages to design a system of objects that form a part-whole hierarchical structure with composite objects formed out of primitive objects. The client does not differentiate between a composite object and a primitive object. The composite hierarchy effectively forms a tree-like hierarchical grouping of objects. From a software engineering perspective, there are at least two problems with the Composite pattern. First, it does not maintain a separation of concerns between the structure of the objects in a system and the objects themselves. The objects that comprise the system contain information about their relationship to other objects. This limits the ability of programmers to reuse the system's structural information. Secondly, there is no mechanism for encapsulating the system as a whole. This makes it difficult to specify and reason about global system properties. This thesis presents two tree components that can be used as alternatives to the Composite design pattern in systems that are traditionally implemented with the pattern. Both components are data structures that can contain arbitrary objects and maintain the structure of those objects as an ordered-tree. Since the components encapsulate only the tree structure, they only need to be specified and verified once, and they are available for black-box reuse. The first component is a traversable tree that maintains a conceptual "cursor" position. Methods are provided for inserting and removing objects at the cursor position, and for moving the cursor throughout the tree. The second component extends the traversable tree. A formal specification for each tree component is presented in the Tako language - a Java-like language with alias avoidance that is designed to facilitate specification and verification. A case study is presented that shows how the indexed tree can be used and reasoned about in an application - a text-based adventure game. Finally, a similar application is developed in Java, once using the composite pattern and once using the indexed tree data structure, and object-oriented metrics are given for both systems.







You are cordially invited to attend Weiping’s PhD Research Defense
Dec. 3, 2pm, 2008, Room 325



Speaker: Weiping He
Advisor: Prof. Ing-Ray Chen


Title: MOBILITY AND SERVICE MANAGEMENT FOR FUTURE ALL-IP BASED WIRELESS NETWORKS

Abstract:


The next generation wireless network will provide not only voice but also data services. With the success of the Internet, it is widely believed that IP will become the foundation of next generation wireless networks. With the help of IETF standardization, IP-based wireless networks can benefit from existing and emerging IP related technologies and services. One key issue is how to provide uninterrupted, reliable and efficient data services to a mobile node (MN) in wireless networks. This dissertation concerns two major system-support mechanisms in future all-IP based wireless networks, namely, mobility management and service management.

Mobility management addresses the issues of how to track and locate a mobile node efficiently. Service management addresses the issues of how to efficiently deliver services to mobile nodes. This dissertation aims to design and analyze integrated mobility and service management schemes for future all-IP based wireless systems. We propose and analyze peruser regional registration schemes extending fromMobile IP Regional Registration (MIP-RR) and Hierarchical MIPv6 (HMIPv6) for integrated mobility and service management with the goal to minimize the network signaling and packet delivery cost in future all-IP based wireless networks. We show that, when given a set of parameters characterizing the operational and workload conditions of a MN, there exists an optimal regional area size for the MN such that the network communication cost is minimized for serving mobility and service management operations of the MN.

If access routers in future all-IP based wireless networks are restricted to perform network layer functions only, we investigate the design of intelligent routers, called dynamic mobility anchor points (DMAPs), to implement per-user regional management in IP wireless networks. These DMAPs are access routers (ARs) chosen by individual MNs to act as regional routers to reduce the signaling overhead for intra-regional movements. The DMAP domain size, i.e., the number of subnets covered by a DMAP, is based on a MNs mobility and service characteristics. A MN optimally determines when and where to launch a DMAP to minimize the network cost in serving the users mobility and service management operations. We show that there exists an optimal DMAP domain size for each individual MN.We also demonstrate that the DMAP design can easily support failure recovery compared with HMIPv6 because of the flexibility of allowing a MN to choose any AR to be the DMAP for mobility and service management.

If access routers are powerful and flexible in future all-IP based networks to perform network-layer and application-layer functions, we propose the use of per-user proxies that can run on access routers. The user proxies can carry service context information such as cached data items and Web processing objects, and perform context-aware functions such as content adaptation for services engaged by the MN to help application executions. We investigate a proxy-based integrated mobility and service management architecture (IMSA) under which a client-side proxy is created on a per-user basis to serve as a gateway between a MN and all services engaged by the MN. Leveraging Mobile IP with route optimization, the proxy runs on an access router and cooperates with the home agent and foreign agent of the MN to maintain the location information of the MN to facilitate data delivery by services engaged by the MN. Further, the proxy optimally determines when to move with the MN so as to minimize the network cost associated with the users mobility and service management operations.

Finally we investigate a proxy-based integrated cache consistency and mobility management scheme called PICMM to support client-server query-based mobile applications, To improve query performance, the MN stores frequently used data in its cache. The MNs proxy receives invalidation reports or updated data objects from application servers, i.e., corresponding nodes (CNs) for cached data objects stored in the MN. If the MN is connected, the proxy will forward invalidation reports or fresh data objects to the MN. If the MN is disconnected, the proxy will store the invalidation reports or fresh data objects, and, once the MN is reconnected, the proxy will forward the latest cache invalidation report or data objects to the MN. We show that there is an optimal service area under which the overall cost including query processing cost and mobility management cost is minimized. To further reduce network traffic, we develop a threshold-based hybrid cache consistency manii agement policy such that whenever a data object is updated at the server, the server sends an invalidation report to the MH through the proxy to invalidate the cached data object only if the size of the data object exceeds the given threshold. Otherwise, the server sends a fresh copy of the data object through the proxy to the MH. We identify the best threshold value that would minimize the overall network traffic incurred due to mobility management, cache consistency management, and query processing, when given a set of parameter values characterizing the operational and workload conditions of the MIP network.

We develop mathematical models to analyze performance characteristics of DMAP, IMSA and PICMM developed in the dissertation research and demonstrate that they outperform existing schemes that do not consider integrated mobility and service management or that use static regional routers to serve all MNs in the system. The analytical results obtained are validated through extensive simulation. We demonstrate the applicability of the design concept of integrated mobility and service management with several mobile client-server applications in MIPv6 systems.






You are cordially invited to attend Okan’s Ph.D. Final Defense
Nov. 17th 2008, 2pm, NVC Room 325

Speaker: Okan Yilmaz
Advisor: Prof. Ing-Ray Chen


Title: A Class of Call Admission Control Algorithms for Resource Management and Reward Optimization for Servicing Multiple QoS Classes in Wireless Networks and Its Applications

Abstract:


Traditional call admission control (CAC) algorithms for resource management in wireless networks which provide connection services to mobile users are designed to satisfy Quality of Service (QoS) constraints of users, such as the blocking probability for new connections and dropping probability for handoff connections. In this research, we develop and analyze a class of CAC algorithms for resource management in wireless networks with the goal not only to satisfy QoS constraints, but also to maximize a value or reward objective function specified by the system. We demonstrate through analytical modeling and simulation validation that the CAC algorithms developed in this dissertation research for resource management can greatly improve the system reward obtainable while satisfying imposed QoS constraints, when compared with existing CAC algorithms designed with QoS satisfaction o

We design these CAC algorithms based on the concept of partitioning or setting proper thresholds to use channel resources for servicing distinct QoS classes. For each CAC algorithm developed, we identify optimal resource management policies in terms of partitioning or threshold settings to use channel resources such that the system reward obtainable is maximized while QoS constraints of service classes are satisfied. By comparing these CAC algorithms head-to-head under identical conditions, we identify the best algorithm to be used at runtime to maximize system reward with QoS guarantees for servicing multiple service classes in wireless networks.

We study solution correctness, solution optimality and solution efficiency of the class of CAC algorithms developed. We ensure solution optimality by comparing optimal solutions obtained with those obtained by ideal CAC algorithms via exhaustive search. We study solution efficiency properties by performing complexity analyses. We ensure solution correctness by simulation validation based on real human mobility data. Further, we analyze the tradeoff between solution optimality vs. solution efficiency and suggest the best CAC algorithm used to best tradeoff solution optimality for solution efficiency, or vice versa, to satisfy the system's solution optimality and solution efficiency requirements.

The major contribution of the dissertation research lies in the development of reward optimization CAC algorithms for resource management in PCS networks to support multiple service classes with distinct QoS constraints. The design principles developed are applicable despite rapidly evolving wireless network technologies since they can be generalized to deal with management of "resources" (e.g., wireless channel bandwidth), "cells" (e.g., cellular networks), "connections" (e.g., service calls with QoS constraints), and "reward optimization" (e.g., revenue optimization in optimal pricing determination) as future wireless service networks must consider both reward maximization and QoS satisfaction when multiplexing limited wireless resources to multiple service classes.

To apply the CAC algorithms developed, we develop an application framework consisting of three stages: workload characterization, call admission control, and application deployment. We demonstrate the applicability with the "optimal pricing" determination application typically performed by service providers, treating "revenue" as a form of the "reward" objective function. Utilizing a simple demand-pricing formula to predict service demands as prices vary, we apply CAC algorithms developed in the dissertation research for reward optimization to determine optimal pricing for multiple service classes in wireless networks such that the overall reward obtained by the system is maximized while QoS constraints of users in multiple service classes are satisfied.






 
 
 
 

You are cordially invited to attend Jin-Hee’s Ph.D. Final Defense
Nov. 12th 2008, 1pm, NVC Room 313

Speaker: Jin-Hee Cho
Advisor: Prof. Ing-Ray Chen


Title: Design and Analysis of QoS-Aware Key Management and Intrusion Detection Protocols for Secure Mobile Group Communications in Wireless Networks

Abstract:


Many mobile applications in wireless networks such as military battlefield, emergency response, mobile commerce, online gaming, and collaborative work are based on the notion of group communications. Designing security protocols for secure group communications in wireless networks faces many technical challenges due to unique characteristics of wireless networks including resource-constrained environments in bandwidth, memory size, battery life and computational power, openness to eavesdropping and security threats, and unreliable communication. Further, for mobile ad hoc networks (MANETs) with no infrastructure support, rapid changes in network topology due to user mobility could cause group merge/partition events to occur dynamically.

While satisfying security requirements is crucial for secure group communications in wireless systems, mobile group applications often have application-specific performance requirements in terms of timeliness, reliability, and system reconfigurability. Often there exists a tradeoff between security versus performance goals since security protocols may introduce undue computational and network overheads which may prevent performance goals from being met.

Unlike traditional security protocols which concern security properties only, in this dissertation research we design and analyze a class of QoS-aware protocols for secure group communications in wireless networks with the goal to satisfy not only security requirements in terms of secrecy, confidentiality, authentication, availability, and data integrity, but also performance requirements in terms of latency, network traffic, response time, and reconfigurability for secure group communication systems (GCSs) in wireless networks. These QoS-aware protocols are adaptive in nature with designs to allow the system to dynamically adjust operational settings, under which both the system’s security and performance requirements can be best satisfied, leveraging the inherent tradeoff between performance versus security goals.

Our contribution has two elements: design and analysis. While our designs mostly derive from existing work, the optimization design principles developed are new to secure GCSs. The analysis methodology developed for the tradeoff analysis of performance versus security of secure group communication protocols is a major contribution. Specifically, the dissertation research has three contributions. First, we propose and analyze efficient, QoS-aware key management protocols for secure group communications in wireless networks to deal with outsider attacks. In order to efficiently reduce the network communication cost caused by rekeying operations (i.e., change a group key), three “threshold-based” periodic batch rekeying protocols are proposed and analyzed. The aim of these protocols is to satisfy application security requirements while minimizing the network communication cost. Instead of individual rekeying, i.e., performing a rekeying operation right after each group join or leave request, these protocols perform batch rekeying periodically. We demonstrate that an optimal rekey interval exists for each protocol that would satisfy an imposed security requirement while minimizing the network communication cost. We further compare these protocols against individual rekeying to identify the best protocol that can minimize the communication cost of rekeying while satisfying application requirements, when given a set of parameter values characterizing the operational and environmental conditions of the system.

Second, we propose and analyze QoS-aware intrusion detection protocols for secure group communications in wireless networks to deal with insider attacks. These protocols explore the tradeoff of security versus performance properties with the goal to determine the best periodic interval for performing intrusion detection. Specifically, we consider a class of intrusion detection protocols including host-based and voting-based IDS protocols for detecting and evicting compromised nodes and examine their effect on MTTSF versus the response time performance metric. Our analysis reveals that there exists an optimal intrusion detection interval under which the MTTSF metric can be best traded off for the response time performance metric, or vice versa. Furthermore, the intrusion detection interval can be dynamically adjusted based on the attacker behaviors to maximize MTTSF while satisfying a system-imposed response time requirement.

Third, we propose and analyze a scalable and efficient region-based group key management protocol for managing mobile groups in MANETs. For scalability and dynamic reconfigurability, we take a region-based approach by which group members are broken into region-based subgroups, and leaders in subgroups securely communicate with each other to agree on a group key in response to membership change and member mobility events. This key management protocol is proposed to identify the optimal regional area size that minimizes the network communication cost while satisfying the application security requirements. Further, it allows mobile groups to react to network partition/merge events for reconfigurability and survivability while still maintaining the design goal of secure group communications in MANETs. Using the proposed region-based group key management, we identify the optimal regional area size that efficiently trades inter-regional communication overhead off for intra-regional communication overhead. We demonstrate its efficiency by comparing it with a no-region GCS, under a set of identified design parameters characterizing network environments and operational conditions of the targeted application. We further investigate the effect of integrating QoS-aware intrusion detection with region-based group key management in MANETs and identify combined optimal settings in terms of the optimal regional size and the optimal intrusion detection interval under which the security and performance properties of the system can be best optimized.

We evaluate the merits of our proposed QoS-aware security protocols for mobile group communications through model-based mathematical analyses with extensive simulation validation. We perform thorough comparative analyses against baseline secure group communication protocols which do not consider security versus performance tradeoffs, including those based on individual rekeying, no intrusion detection, and/or no-region designs. The results obtained show that our proposed QoS-aware security protocols outperform these baseline algorithms.






You are cordially invited to attend Okan’s Ph.D. Research Defense
August 14th 2008, 2pm, NVC Room 324

Speaker: Okan Yilmaz
Advisor: Prof. Ing-Ray Chen


Title: A Class of Call Admission Control Algorithms for Resource Management and Reward Optimization for Servicing Multiple QoS Classes in Wireless Networks and Its Applications

Abstract:


Traditional call admission control (CAC) algorithms for resource management in wireless networks which provide connection services to mobile users are designed to satisfy Quality of Service (QoS) requirements of users, such as the blocking probability for new connections and dropping probability for handoff connections. In this research, we develop and analyze a class of CAC algorithms for resource management in wireless networks with the goal not only to satisfy QoS requirements, but also to maximize a value or reward objective function specified by the system. We demonstrate through analytical modeling and simulation validation that the CAC algorithms developed in this dissertation research for resource management can greatly improve the system reward obtainable while satisfying imposed QoS requirements, when compared with existing CAC algorithms designed with QoS satisfaction only.

We design these CAC algorithms based on the concept of partitioning or setting proper thresholds to use channel resources for servicing distinct QoS classes. For each CAC algorithm developed, we identify optimal resource management policies in terms of partitioning or threshold settings to use channel resources such that the system reward obtainable is maximized while QoS requirements of service classes are satisfied. By comparing these CAC algorithms head-to-head under identical conditions, we identify the best algorithm to be used at runtime to maximize system reward with QoS guarantees for servicing multiple service classes in wireless networks.

We study solution correctness, solution optimality and solution efficiency of the class of CAC algorithms developed. We ensure solution optimality by comparing optimal solutions obtained with those obtained by ideal CAC algorithms via exhaustive search. We study solution efficiency properties by performing complexity analyses. We ensure solution correctness by simulation validation based on real human mobility data. Further, we analyze the tradeoff between solution optimality vs. solution efficiency and suggest the best CAC algorithm used to best tradeoff solution optimality for solution efficiency, or vice versa, to satisfy the system's solution optimality and solution efficiency requirements.

The major contribution of the dissertation research lies in the development of reward optimization CAC algorithms for resource management in PCS networks to support multiple service classes with distinct QoS requirements. The design principles developed are applicable despite rapidly evolving wireless network technologies since they can be generalized to deal with management of "resources" (e.g., wireless channel bandwidth), "cells" (e.g., cellular networks), "connections" (e.g., service calls with QoS requirements), and "reward optimization" (e.g., revenue optimization in optimal pricing determination) as future wireless service networks must consider both reward maximization and QoS satisfaction when multiplexing limited wireless resources to multiple service classes.

To apply the CAC algorithms developed, we develop an application framework consisting of three stages: workload characterization, call admission control, and application deployment. We demonstrate the applicability with the "optimal pricing" determination application typically performed by service providers, treating "revenue" as a form of the "reward" objective function. Utilizing a simple demand-pricing formula to predict service demands as prices vary, we apply CAC algorithms developed in the dissertation research for reward optimization to determine optimal pricing for multiple service classes in wireless networks such that the overall reward obtained by the system is maximized while QoS requirements of users in multiple service classes are satisfied.






 
 
 
 

You are cordially invited to attend Jin-Hee’s Ph.D. Research Defense
May 29th 2008, 2pm, NVC Room 320

Speaker: Jin-Hee Cho
Advisor: Prof. Ing-Ray Chen


Title: Design and Analysis of QoS-Aware Key Management and Intrusion Detection Protocols for Secure Mobile Group Communications in Wireless Networks

Abstract:


Many mobile applications in wireless networks such as military battlefield, emergency response, mobile commerce, online gaming, and collaborative work are based on the notion of group communications. Designing security protocols for secure group communications in wireless networks faces many technical challenges due to unique characteristics of wireless networks including resource-constrained environments in bandwidth, memory size, battery life and computational power, openness to eavesdropping and security threats, and unreliable communication. Further, for mobile ad hoc networks (MANETs) with no infrastructure support, rapid changes in topology due to user mobility could cause group merge/partition events to occur dynamically.

While satisfying security requirements is crucial for secure group communications in wireless systems, mobile group applications often have application-specific performance requirements in terms of timeliness, reliability, and system reconfigurability. Often there exists a tradeoff between security vs. performance goals since security protocols may introduce undue computational and network overheads which may prevent performance goals from being met.

Unlike traditional security protocols which concern security properties only, in this dissertation research we design and analyze a class of QoS-aware protocols for secure group communications in wireless networks with the goal to satisfy not only security requirements in terms of secrecy, confidentiality, authentication, availability, and data integrity, but also performance requirements in terms of latency, network traffic, response time, and reconfigurability for secure group communication systems (GCSs) in wireless networks. These QoS-aware protocols are adaptive in nature with designs to allow the system to dynamically adjust operational settings, under which both the system’s security and performance requirements can be best satisfied, leveraging the inherent tradeoff between performance vs. security goals.








 
 
 
 

You are cordially invited to attend Matthew’s MS Defense
April 24th 2008, 2pm, NVC Room 322

Speaker: Matthew C. Makai
Co-Advisor: Prof. Ing-Ray Chen and Prof. Greg Kulczycki


Title: Incorporating Design Knowledge into Genetic Algorithm-based White-Box Software Test Case Generators

Abstract:


This thesis shows how design knowledge can be extracted from Unified Modeling Language sequence diagrams and incorporated into genetic algorithm-based automated test case generators to increase the coverage of their test cases. Automated generation of test data through evolutionary testing was proven feasible in prior research studies. In those previous investigations, the metrics used for determining the test generation method effectiveness were the percentages of testing statement and branch code coverage achieved. However, the code coverage realized in those preceding studies often converged at suboptimal percentages due to a lack of guidance in conditional statements. This study examines a tool known as the Evolutionary Test Case Generator, or ETCG (pronounced "e-tee-see-gee"), which provides an improved method for automatically producing test case suites using genetic algorithms. Test case production is accomplished by a novel method for incorporating design knowledge into the evolutionary test case generation process. In this study, common Unified Modeling Language sequence diagrams provide design knowledge to direct the heuristic search process and facilitate the production of test cases. The design knowledge, measured by the number of sequence diagrams associated with the source code incorporated into the generation process, provides guidance to the searches for apposite test cases. When the generator uses design knowledge, the resulting test cases converge at higher code coverage percentages that are unattainable when the design knowledge is not utilized.






You are cordially invited to attend Ahn’s PhD Final Defense
April 17, 2008, 1pm, NVC Room 320



Speaker: Anh Phan Speer
Advisor: Prof. Ing-Ray Chen


Title: Design and Analysis of Adaptive Fault Tolerant QoS Control Algorithms for Query Processing in Wireless Sensor Networks

Abstract:


Wireless sensor networks (WSNs) present several unique characteristics such as resource-constrained sensors, random deployment, and data-centric communication protocols. These characteristics pose unprecedented challenges in the area of query processing in WSNs. This dissertation presents the design and validation of adaptive fault tolerant QoS control algorithms with the objective to achieve the desired quality of service (QoS) requirements and maximize the system lifetime in query-based WSNs.

Data sensing and retrieval in WSNs have a great applicability in military, environmental, medical, home and commercial applications. In query-based WSNs, a user would issue a query with QoS requirements in terms of reliability and timeliness, and expect a correct response to be returned within the deadline. Satisfying these QoS requirements requires that fault tolerance mechanisms through redundancy be used, which may cause the energy of the system to deplete quickly. We analyze the effect of redundancy on the mean time to failure (MTTF) of query-based cluster-structured WSNs, defined as the mean number of queries that a WSN is able to answer correctly until it fails due to channel faults, sensor faults, or sensor energy depletion. We show that a tradeoff exists between redundancy and MTTF. Furthermore, an optimal redundancy level exists such that the MTTF of the system is maximized.

We develop a hop-by-hop data delivery (HHDD) mechanism and an Adaptive Fault Tolerant Quality of Service Control (AFTQC) algorithm in which we utilize "source" and "path" redundancies with the goal to satisfy application QoS requirements while maximizing the lifetime of WSNs. We also compare and contrast AFTQC without acknowledgment vs. AFTQC with acknowledgement and identify conditions under which AFTQC should couple with acknowledgement to maximize the system MTTF.

To deal with network dynamics, we investigate proactive and reactive methods to dynamically collect channel and delay conditions to determine the optimal redundancy level at runtime. AFTQC can adapt to network dynamics that cause changes to the node density, residual energy, sensor failure probability, and radio range due to energy consumption, node failures, and change of node connectivity. Further, AFTQC can deal with software faults, concurrent query processing with distinct QoS requirements, and data aggregation.

We compare our design with a baseline design without redundancy based on acknowledgement for data transmission and geographical routing for relaying packets to demonstrate the feasibility. We validate analytical results with extensive simulation studies. When given QoS requirements of queries in terms of reliability and timeliness, our AFTQC design allows optimal "source" and "path" redundancies to be identified and applied dynamically in response to network dynamics such that not only query QoS requirements are satisfied, as long as adequate resources are available, but also the lifetime of the system is maximized.





You are cordially invited to attend Ahn’s PhD Research Defense
Dec. 20 2007, 11am, NVC Room 320



Speaker: Anh Phan Speer
Advisor: Prof. Ing-Ray Chen


Title: Design and Analysis of Adaptive Fault Tolerant QoS Control Algorithms for Query Processing in Wireless Sensor Networks

Abstract:


Wireless sensor networks (WSNs) present several unique characteristics such as resource-constrained sensors, random deployment, and data-centric communication protocols. These characteristics pose unprecedented challenges in the area of query processing in WSNs. This dissertation presents the design and validation of adaptive fault tolerant QoS control algorithms with the objective to achieve the desired quality of service (QoS) requirements and maximize the system lifetime in query-based WSNs. Data sensing and retrieval in WSNs have a great applicability in military, environmental, medical, home and commercial applications. In query-based WSNs, a user would issue a query with QoS requirements in terms of reliability and timeliness, and expect a correct response to be returned within the deadline. Satisfying these QoS requirements requires that fault tolerance mechanisms through redundancy be used, which may cause the energy of the system to deplete quickly. We analyze the effect of redundancy on the mean time to failure (MTTF) of query-based cluster-structured WSNs, defined as the mean number of queries that a WSN is able to answer correctly until it fails due to channel faults, sensor faults, or sensor energy depletion. We show that a tradeoff exists between redundancy and MTTF. Furthermore, an optimal redundancy level exists such that the MTTF of the system is maximized. We develop a hop-by-hop data delivery (HHDD) mechanism and an Adaptive Fault Tolerant Quality of Service Control (AFTQC) algorithm in which we utilize "source" and "path" redundancies with the goal to satisfy application QoS requirements while maximizing the lifetime of WSNs. We also compare and contrast AFTQC without acknowledgment vs. AFTQC with acknowledgement and identify conditions under which AFTQC should couple with acknowledgement to maximize the system MTTF. To deal with network dynamics, we investigate proactive and reactive methods to dynamically collect channel and delay conditions to determine the optimal redundancy level at runtime. AFTQC can adapt to network dynamics that cause changes to the node density, residual energy, sensor failure probability, and radio range due to energy consumption, node failures, and change of node connectivity. Further, AFTQC can deal with software faults, concurrent query processing with distinct QoS requirements, and data aggregation. We compare our design with a baseline design without redundancy based on acknowledgement for data transmission and geographical routing for relaying packets to demonstrate the feasibility. We validate analytical results with extensive simulation studies. When given QoS requirements of queries in terms of reliability and timeliness, our results allow optimal "source" and "path" redundancies to be identified by AFTQC such that not only QoS requirements are satisfied, but also the lifetime of the system is prolonged. Finally AFTQC is demonstrated to maximize the MTTF of the system despite the presence of network dynamics.







You are cordially invited to attend Abdelmounaam’s PhD Final Defense
November 29th 2007, 10am, NVC Room 106



Speaker: Abdelmounaam Rezgui
Advisor: Prof. Mohamed Eltoweissy


Title: Service-Oriented Sensor-Actuator Networks

Abstract:


For decades since their inception, sensor-actuator networks (SANETs) have been closed networks owned, maintained, and used by a single party, e.g., a government agency, a research institution, or a private company. Typically, the network is designed and deployed to serve one or a few applications with a specic set of characteristics. In subsequent years, the need to decouple SANETs from the applications using them led to the emergence of generic SANETs, an alternative design model where an application-independent query system is deployed on each node of the SANET. In this model, the query system is designed to answer queries from any application. Both application-specic and generic SANET architectures are inherently inadequate to support the next-generation of open, interoperable, pervasive, multi-purpose, Web-accessible SANETs. Indeed, application-specic SANETs provide limited reusability, are not cost eective, and may require extensive reprogramming eorts when new applications need to use the network. Generic SANETs usually require that a sizeable code be deployed on the nodes regardless of the specic requirements of the application at hand. More importantly, they may not be optimized to fully exploit the specic characteristics and query patterns of a given application. We argue that the next-generation of SANETs require customizable architectures where SANETs may consist of nodes that have heterogeneous hardware and software components and are managed by autonomous entities. Customizable SANETs would provide developers the ability to select individual software components from several SANETs that are already deployed and integrate these components in new applications. These SANETs must expose their capabilities to developers at an adequate level of abstraction and, yet, achieve high eciency, scalability and reusability. In this dissertation, we propose service-oriented SANETs (SOSANETs) as a novel approach for building customizable SANETs. In SOSANETs, nodes expose their capabilities to applications in the form of service proles. A node's service prole consists of a set of services (i.e. sensing and actuation capabilities) that it provides and the quality of service (QoS) parameters associated with those services (delay, accuracy, freshness, etc.). Services are lightweight code units deployed directly on top of the operating system of nodes. SOSANETs provide the benets of both application-specic SANETs (e.g., energy eciency, scalability) and generic SANETs (e.g., reusability), and avoid most of their limitations. Developing SOSANETs entails, in particular, three major research challenges: (i) defining a query model and a system architecture to support that model, (ii) developing new routing protocols, and (iii) developing new query processing techniques. In this research, we address each of these challenges and demonstrate SOSANETs's potential in supporting the next-generation of sensor infrastructures. Our contributions may be summarized as follows:

1. Service-Oriented Query Model and Architecture for SANETs: We introduce service-oriented sensor-actuator networks as a new paradigm for building the next generation of open, interoperable, customizable sensor-actuator networks. We dene a query model for SOSANETs and propose an architecture that supports that model. The proposed query model oers a simple, uniform query interface whereby applications specify sensing and actuation queries independently from any specic deployment of the underlying SOSANET.

2. Routing Protocols: We developed RACER (Reliable Adaptable serviCe-driven Efficient Routing), a routing protocol suite for SOSANETs. RACER consists, essentially, of three routing protocols, namely, SARP (Service-Aware Routing Protocol), TARP (Trust-Aware Routing Protocol), and CARP (Context-Aware Routing Protocol). RACER uses an ecient service-aware routing approach that aggressively reduces downstream trac (from the sink to the network's nodes) by translating service proles into ecient paths for queries. To support QoS, RACER dynamically adapts each node's routing behavior and service prole according to the current context of that node, i.e. number of pending queries and number and type of messages to be routed. Finally, RACER improves end-to-end reliability through a scalable reputation-based approach in which each node is able to locally estimate the next hop of the most reliable path to the sink.

3. Query Processing and Optimization: We introduce a set of query optimization techniques that contribute to the ecient execution of queries. These techniques include service-driven sleep scheduling, service-based multi-query processing, and multi-event detection.

4. Implementation and Evaluation: To validate our work, we implemented TinySOA, a prototype SOSANET built on top of TinyOS 1.1.15 with RACER as its routing mechanism. A design approach that has proved its merit in wireless networks is cross- layer optimization. Given the promise of this design approach and the advantages of layered design, we designed and implemented TinySOA as a set of layers with a loose interaction model that enables several cross-layer optimization options. We conducted an evaluation of TinySOA that included a comparison with TinyDB, an established query processing system for sensor networks. The obtained empirical results show that TinySOA achieves signicant improvements on many aspects including energy consumption, scalability, reliability and response time.

Sensor-actuator networks are an integral component in the vision of ubiquitous service environments. As they mature, SANETs will be used in a wide spectrum of applications. This will transform SANETs into a "utility" that will support much of the societal activities. Providers will deliver this utility to "consumers" as a service. To SANET providers and consumers, the challenges will be the same: improving eciency, reliability, scalability, and interoperability while reducing development and maintenance cost. This will translate into a host of new requirements. Current architectures are inherently unable to support those requirements. They also are not able to properly exploit the benets of recent and future advances in related technologies. The architecture and protocols proposed in this dissertation are a fundamental departure from all existing approaches for SANET architectures and protocols. Theoretical and empirical results show that a service-oriented architecture design is a viable candidate to support the requirements of tomorrow's sensor-actuator networks. We regard our work as a milestone in demonstrating the potential of SOSANETs. We expect its impact to be signicant on the design, development, and deployment of future SANETs. This impact will span several aspects including cost, development time, eciency, interoperability, and scalability. In fact, our work readily applies to any SANET-like wireless networks where the main characteristics are nodes with limited energy supply and lossy communication links. We anticipate that this work will foster research that would reformulate and address many of the problems encountered in developing and deploying today's sensor systems.








You are cordially invited to attend Xu’s PhD Final Defense
November 26th 2007, 5pm, NVC Room 111



Speaker: Xu Yang
Advisor: Prof. Athman Bouguettaya


Title: Multi-channel Mobile Access to Web Services

Abstract:


To support wireless-oriented services, a new generation of Web services called Mobile services (M-services) has emerged. M-services provide mobile users access to services through wireless networks. One of the important issues in M-service environment is how to discover and access M-services efficiently. In this dissertation, we propose time and power efficient access methods for M-services. We focus on methods for accessing broadcast based M-services from multiple wireless channels. We first discuss efficient access methods in data-oriented wireless broadcast sys- tems. We then discuss how to extend current wireless broadcast systems to support simple M-services. We present a novel infrastructure that provides a multi-channel broadcast framework for mobile users to effectively discover and access composite M-services. Multi-channel algorithms are proposed for efficiently accessing composite services. We define a few semantics that have impact on access efficiency in the pro- posed infrastructure. We discuss semantic access to composite services. Broadcast channel organizations suitable for discovering and accessing composite services are proposed. We also derive analytical models for these channel organizations. To provide practical study for the proposed infrastructure and access methods, a testbed is developed for simulating accessing M-services in a broadcast-based environment. Extensive experiments have been conducted to study the proposed access methods and broadcast channel organizations. The experimental results are presented and discussed.








You are cordially invited to attend Neha’s MS Defense
August 10th 2007, 2pm, NVC Room 314



Speaker: Neha Khedekar
Advisor: Prof. Gregory W. Kulczycki


Title: Empirical Analysis of Value and Reference Semantics

Abstract:


In this thesis, we attempt to measure the impact of reference semantics on programming and reasoning. There is a lot of anecdotal evidence that references and aliasing complicate both formal and informal reasoning, but there is a lack of empirical data on the topic. In this thesis, we have designed a survey that is used to compare how well programmers perform under different programming paradigms. Two of the programming paradigms studied in the survey, copying and swapping, use value semantics, while the third, referencecopying, uses reference semantics. We have given the survey to over 25 people who have various levels of Java programming experience. The results of the survey seem to support the anecdotal evidence that programming with value semantics is easier than programming with reference semantics.








You are cordially invited to attend Amrinder’s MS Defense
August 9th 2007, 2pm, NVC Room 325



Speaker: Amrinder Singh
Advisor: Prof. Gregory W. Kulczycki


Title: A component-based approach to proving the correctness of the Schorr-Waite algorithm

Abstract:


This thesis presents a component-based approach to proving the correctness of programs involving pointers. Unlike previous work, our component-based approach supports modular reasoning, which is essential to the scalability of systems. Specifically, we specify the behavior of a graph-marking algorithm known as the Schorr-Waite algorithm, implement it using a component that captures the behavior and performance benefits of pointers, and prove that the implementation is correct with respect to the specification. We use the Resolve language in our example, which is an integrated programming and specification language that supports modular reasoning. The behavior of the algorithm is fully specified using custom definitions, pre- and post-conditions, and a complex loop invariant. Additional operations for the Resolve pointer component are introduced that preserve the accessibility of a system. These operations are used in the implementation of the algorithm. They simplify the proof of correctness and make the code shorter.








You are cordially invited to attend Jianghui’s PhD Research Defense
July 19 2007, 5pm, NVC Room 111



Speaker: Jianghui Ying
Advisor: Prof. Denis Gracanin


Title: Support for Subjective Views in Collaborative Virtual Environments

Abstract:


Collaborative Virtual Environments (CVEs) use a shared virtual world to support interactions and collaboration among users. The majority of CVEs provide a highly objective virtual environment. That is, each user is presented with the same virtual world in the same way, albeit from diRerent viewpoints. This is partly due to the fact that multi-user Virtual Reality (VR) systems have evolved from single-user systems that have been extended to support a number of users. The development trend of VR systems is similar to the way that groupware systems evolved from 2D single-user systems that simply replicated the single-user interface to multiple users. WYSIWIS (What You See Is What I See) is the foundational abstraction that guided the multi-user interface design of early groupware systems. WYSIWIS is critical for collaboration. However, some research has indicated that the strict objectivity is too inXexible. In some cases it may even hinder collaboration if all users are forced to work on the same representation without the capability to tailor the view to meet their own needs. This has led to the development of the Relaxed-WYSIWIS 2D user interface design. Just as the strict WYSIWIS proved too limited in 2D shared user interface, there is a need to extend current multi-user virtual environments to support subjectivity. The concept of "Subjective View" is introduced to give users ability to control the presentation of the virtual world to best suit their needs. In a CVE interface, subjectivity is a dicult issue. It can provide great benet to users, however, it might also pose problems for users' cooperation if used inappropriately. The goal of this research is to explore the approach of supporting appropriate subjective views in Collaborative Virtual Environments (CVEs). The hypothesis of this research is that, under certain conditions, subjective views can improve user and task performance over certain tasks in CVEs compared to the corresponding objective view. Our approach is based on the assumption that in a CVE, only a subset of information, core information, about the environment is needed to support collaboration. The rest of the informa- tion, auxiliary information, is not important for collaboration tasks. Auxiliary information can be presented in diRerent ways without impacting on the collaboration tasks. To build an eRective Subjective Views CVE (SVCVE) system, we need to maintain consistent awareness among collaborative users under subjective views. To achieve that, we look at a CVE as composed of two layers: the syntax layer, which is the cooperative interface specic information, and the semantics layer, which is the underlying application functionality. In a subjective environment, what users see from the syntax layer could be very diRerent. However, from the semantics level, all users should get the same or nearly the same knowledge. We believe the semantics associated with the CVE is the foundation to maintain consistency. The CVE semantic is modeled using a formal method - Petri net. Petri net is a promising formal model. It is a state-oriented formal specication notation. The Petri net model provides a foundation for our work. In order to support a general way of building SVCVE systems, we presented a framework sup- porting the process of going from analyzing initial 3D content describing the virtual environment to creating corresponding subjective views for the CVE. We discussed a general subjective views CVE model and a corresponding Petri net based hierarchical model to describe the relationship among diRerent views in SVCVE system. Based on that, a SVCVE system implementation framework is proposed. In order to illustrate the process and the implementation framework, we implemented a prototype system and used it for case studies. Furthermore, we discussed how to analyze SVCVE system characteristics using Petri net inherent properties such as reachability, reinitiability, boundedness, etc. We also performed a pilot study using a simple case study to evaluate user and task performance in diRerent subjective views. when providing views with suitable degree of subjectivity in CVE, user would have better performance in certain collaborative tasks. In future work, in order to prove the above hypothesis, we will evaluate the user and task performance by applying subjective views under two conditions: symmetric condition (same resources such as platforms, input/output devices, etc.) and asymmetric condition (diRerent resources). Under symmetric condition, we will focus on how users' personal preferences will help to improve the performance while maintaining semantic consistency. Under asymmetric condition, we will focus on leveraging subjective views to minimize the semantic inconsistency caused by system diRerence in order to better support eRective collaboration.








You are cordially invited to attend Xu’s PhD Research Defense
June 14, 5pm, 2007, Room 103



Speaker: Xu Yang
Advisor: Prof. Athman Bouguettaya


Title: Multi-channel Mobile Access to Web Services

Abstract:


To support wireless-oriented services, a new generation of Web services called Mobile services (M-services) has emerged. M-services provide mobile users access to services through wireless networks. One of the important issues in M-service environment is how to discover and access M-services efficiently. The aim of our research is to investigate time and power efficient access methods to M-services. We focus on methods for accessing broadcast based M-services from multiple wireless channels. We also aim to provide a generic broadcast based M-service infrastructure for delivering services to mobile users. In our preliminary work, we have discussed efficient data access methods and how to organize and access simple M-services in a broadcast based environment. In this report, we focus on efficient access to composite services. We first demonstrate the major challenges of efficiently accessing broadcast-based composite services. Then we present an enhanced M-services infrastructure that provides a framework for mobile users to effectively discover and access broadcast-based composite M-services. The proposed infrastructure uses multiple broadcast channels to deliver required wireless information. Multi-channel algorithms are proposed for efficiently accessing composite services. We define a few semantics that have impact on access efficiency in the proposed infrastructure. We discuss semantic access to composite services. Broadcast channel organizations suitable for discovering and accessing composite services are proposed. Analytical models for these channel organizations are derived. To provide practical study for the proposed infrastructure and access methods, a testbed is implemented for simulating accessing composite services in a broadcast- based environment. Extensive experiments have been conducted to study the proposed semantic access method and broadcast channel organizations. The experimental results are presented and discussed.








You are cordially invited to attend Jin-Hee’s PhD Preliminary Exam defense
May 9, 10am, 2007. Room 320



Speaker: Jin-Hee Cho
Advisor: Prof. Ing-Ray Chen


Title: Design and Analysis of QoS-Aware Key Management and Intrusion Detection Protocols for Secure Mobile Group Communications in Wireless Networks

Abstract:


Many mobile applications in wireless networks such as military battlefield, emergency response, mobile commerce, online gaming, and collaborative work are based the notion of group communications. Designing security protocols for secure group communications in wireless networks faces many technical challenges due to unique characteristics of wireless networks including resource-constrained environments in bandwidth, memory size, battery life, computational power, etc., openness to eavesdropping and security threats, unreliable communication, and, for mobile ad hoc networks with no infrastructure support, rapid changes in topology due to user mobility which can cause group merge/partition events to occur dynamically.
While satisfying security requirements is crucial for secure group communications in wireless systems, mobile group applications often have application-specific performance requirements in terms of timeliness, reliability, and system reconfigurability. Often there exists a tradeoff between security vs. performance goals since security protocols may introduce undue computational and network overheads which may prevent performance goals from being met.
Unlike traditional security protocols which concern security properties only, in this dissertation research we propose and analyze a class of QoS-aware protocols for secure group communications in wireless networks with the goal to satisfy not only security requirements in terms of secrecy, confidentiality, authentication, availability, and data integrity, but also performance requirements in terms of latency, network traffic, response time, and reconfigurability for secure group communication systems in wireless networks. These QoS-aware protocols are adaptive in nature with designs to allow the system to dynamically adjust operational settings, under which both the system’s security and performance requirements can be best satisfied, leveraging the inherent tradeoff between performance vs. security goals.
This dissertation research has three contributions. First, we propose and analyze efficient, QoS-aware key management protocols for secure group communications in wireless networks to deal with outsider attacks. In order to efficiently reduce the network communication cost caused by rekeying operations (to change a group key), three “threshold-based” periodic batch rekeying protocols are proposed and analyzed. The aim of these protocols is to satisfy application security requirements while minimizing the network communication cost. Instead of individual rekeying, i.e., performing a rekeying operation right after each group join or leave request, these protocols perform batch rekeying periodically. We demonstrate that an optimal rekey interval exists for each protocol that would satisfy an imposed security requirement while minimizing the network communication cost. We further compare these protocols against individual rekeying to identify the best protocol that can minimize the communication cost of rekeying while satisfying application requirements when given a set of parameter values characterizing the operational and environmental conditions of the system. We report results for the case in which a centralized key server exists in wireless networks. Future work will remove this restriction and investigate designs to apply these threshold-based periodic batch rekeying protocols to infrastructure-less mobile ad hoc networks without a centralized key server.
Second, we propose and analyze QoS-aware intrusion detection protocols for secure group communications in wireless networks to deal with insider attacks. These protocols explore the tradeoff of security vs. performance properties with the goal to determine the best periodic interval for performing intrusion detection. Specifically, we consider a class of intrusion detection protocols including host-based and voting-based protocols for detecting and evicting compromised nodes and examine their effect on the mean time to security failure (MTTSF) vs. the response time performance metric. Our analysis reveals that there exists an optimal intrusion detection interval under which the MTTSF metric can be best traded off for the response time performance metric, or vice versa. Furthermore, the intrusion detection interval can be dynamically adjusted based on the attacker behaviors to maximize MTTSF while satisfying a system-imposed response time requirement. We report results for the case in which all nodes are covered by a single-hop peer-to-peer wireless network. Future work involves extending the design and analysis to the case in which mobile nodes communicate with each other through multi-hop in mobile ad hoc networks.
Third, we propose and analyze a scalable and efficient region-based group key management protocol for managing mobile groups in mobile ad hoc networks. For scalability and dynamic reconfigurability, we take a regionbased approach by which group members are broken into region-based subgroups, and leaders in subgroups securely communicate with each other to agree on a group key in response to membership change and member mobility events. This key management protocol is proposed to identify the optimal size of a region that minimizes the network communication cost while satisfying the application security requirements. Further, it allows mobile groups to react to network partitioning/merging events for reconfigurability and survivability while still maintaining the design goal of secure group communications in mobile ad hoc networks.
We propose to extend and integrate our research in periodic batch rekeying (for dealing with outsider attacks) and distributed intrusion detection (for dealing with insider attacks) into the region-based group key management protocol for secure group communications in mobile ad hoc networks. The target mobile group communication system will be built upon our proposed two-level hierarchical key management structure for scalability, reconfigurability and efficiency, as well as for adaptability to allow the system to dynamically choose the best operational setting in response to runtime network conditions to trade performance off for security, or vice versa, to best satisfy the application imposed performance and security requirements. We propose to perform comparative analyses against secure group communication protocols that do not consider security vs. performance tradeoffs, including those based on individual rekeying, static intrusion detection, and/or no-region designs. We propose to evaluate the merits of our proposed QoS-aware security protocols for mobile group communications through modelbased mathematical analyses with simulation validation.








You are cordially invited to attend Yilmaz’s PhD Preliminary Exam defense
March 21, 2pm, 2007. Room 207



Speaker: Okan Yilmaz
Advisor: Prof. Ing-Ray Chen


Title: A Framework for Resource and Pricing Management for Revenue Optimization with QoS Guarantees for Multiple Service Classes in Wireless Networks

Abstract:


We develop a framework for resource and pricing management for revenue optimization with Quality of Service (QoS) guarantees in personal communication service (PCS) wireless networks that provide multiple service classes to roaming mobile users.
The framework proposed in the dissertation research consists of three parts. The first part is the development of a workload characterization algorithm for calculating the arrival and departure rates of multiple multimedia service class from data statistically collected by individual mobile users and the system. The workload characterization algorithm provides information for the system to make resource management decisions to admit or reject service calls made by mobile users with the objective of maximizing revenue with QoS guarantees.
he second part of the research is the development of a class of admission control algorithms that make acceptance decisions to new and handoff calls of multiple service classes to satisfy specified QoS constraints in terms of the dropping probability of handoff calls and the blocking probability of new calls. We utilize these call admission control algorithms to determine the maximum revenue obtainable by each cell of the PCS system while satisfying the QoS constraints, when given predetermined prices for multiple service classes.
The third part of the research is to develop and analyze a constraint-based search algorithm to efficiently determine optimal pricing that maximizes the system revenue with QoS guarantees. This algorithm first determines a number of possible price values for each service class, thus creating a search space consisting of possible price combinations of all multimedia services. Then it explores the search space guided by QoS constraints and a service demand-price correlation scheme to effectively reduce the search complexity without degrading the solution optimality. The constraint-based algorithm is shown to effectively search for optimal pricing that maximizes the system revenue with QoS guarantees in all cells of the PCS system.
We propose to use mathematical modeling and analysis methods to obtain analytical results and compare the proposed set of algorithms developed for the framework with baseline algorithms in the literature. We plan to use simulation to validate analytical results.







You are cordially invited to attend Weiping’s PhD Preliminary Exam defense
Dec. 12, 2pm, 2006. Room 324



Speaker: Weiping He
Advisor: Prof. Ing-Ray Chen


Title: MOBILITY AND SERVICE MANAGEMENT FOR FUTURE ALL-IP BASED WIRELESS NETWORKS

Abstract:


The next generation wireless network will provide not only voice but also data services. With the success of the Internet, it is widely believed that IP will become the foundation of next generation wireless networks. With the help of IETF standardization, IP-based wireless networks can benefit from existing and emerging IP related technologies and services. One key issue is how to provide uninterrupted, reliable and efficient data services to a mobile node (MN) in wireless networks. This dissertation concerns two major system-support mechanisms in future all-IP based wireless networks, namely, mobility management and service management.
Mobility management addresses the issues of how to track and locate a mobile node efficiently. Service management addresses the issues of how to efficiently deliver services to mobile nodes. This dissertation aims to design and analyze integrated mobility and service management schemes for future all-IP based wireless systems. We propose and analyze per-user regional registration schemes for integrated mobility and service management with the goal to minimize the network signaling and packet delivery cost in future all-IP based wireless networks. We show that, when given a set of parameters characterizing the operational and workload conditions of a MN, there exists an optimal regional area size for the MN such that the network communication cost is minimized for serving mobility and service management operations of the MN.
If access routers in future all-IP based wireless networks are restricted to perform network layer functions only, we investigate the design of intelligent routers, called dynamic mobility anchor points (DMAPs), to implement per-user regional management in IP wireless networks. These DMAPs are access routers (ARs) chosen by individual MNs to act as regional routers to reduce the signaling overhead for intra-regional movements. The DMAP domain size, i.e., the number of subnets covered by a DMAP, is based on a MN's mobility and service characteristics. A MN optimally determines when and where to launch a DMAP to minimize the network cost in serving the user's mobility and service management operations. We show that there exists an optimal DMAP domain size for each individual MN.
If access routers are powerful and flexible in future all-IP based networks to perform network-layer and application-layer functions, we propose the use of per-user proxies that can run on access routers. The user proxies can carry service context information such as cached data items and Web processing objects, and perform context-aware functions such as content adaptation for services engaged by the MN to help application executions. Under the proxy-based regional management scheme, a client-side proxy is created on a per-user basis to serve as a gateway between a MN and all services engaged by the MN. Leveraging Mobile IP with route optimization, the proxy runs on a foreign agent/access router and cooperates with the home agent and foreign agent/access router of the MN to maintain the location information of the MN, in order to facilitate data delivery by services engaged by the MN. Further, the proxy optimally determines when to move with the MN so as to minimize the network cost associated with the user's mobility and service management operations.
The proxy-based scheme supports query processing mobile applications. To improve query performance, the MN stores frequently used data in its cache. The MN's proxy receives invalidation reports or updated data objects from application servers, i.e., corresponding nodes (CN) for cached data objects stored in the MN. If the MN is connected, the proxy will forward invalidation reports or fresh data objects to the MN. If the MN is disconnected, the proxy will store the invalidation reports or fresh data objects, and, once the MN is reconnected, the proxy will forward the latest cache invalidation report or data objects to the MN. We show that there is an optimal ``service area'' under which the overall cost including query processing cost and mobility management cost is minimized.
We demonstrate that our proposed per-user regional management scheme outperforms basic Mobile IPv6, Mobile IPv6 Regional Registration, and Hierarchical Mobile IPv6 that do not consider integrated mobility and service management and that use static regional routers to serve all MNs in the system. We will develop a simulation model based on ns2 to validate analytical results. We will also investigate mobile applications to which the proposed integrated mobility and service management scheme can be applied in Mobile IP systems.



 


 
 
 
 

You are cordially invited to attend Kou’s PhD Final Exam Defense
Nov. 29, 10am, 2006. Room 103



Speaker: Yufeng Kou
Advisor: Prof. Chang-Tien Lu


Title: Abnormal Pattern Recognition in Spatial Data

Abstract:


In the recent years, abnormal spatial pattern recognition has received a great deal of attention from both industry and academia, and has become an important branch of data mining. Abnormal spatial patterns, or spatial outliers, are those observations whose characteristics are markedly dierent from their spatial neighbors. The identication of spatial outliers can be used to reveal hidden but valuable knowledge in many applications. For example, it can help locate extreme meteorological events such as tornadoes and hurricanes, identify aberrant genes or tumor cells, discover highway trac congestion points, pinpoint military targets in satellite images, determine possible locations of oil reservoirs, and detect water pollution incidents. Numerous traditional outlier detection methods have been developed, but they cannot be directly applied to spatial data in order to extract abnormal patterns. Traditional outlier detection mainly focuses on "global comparison" and identies deviations from the remainder of the entire data set. In contrast, spatial outlier detection concentrates on discovering neighborhood instabilities that break the spatial continuity. In recent years, a number of techniques have been proposed for spatial outlier detection. However, they have the following limitations. First, most of them focus primarily on single-attribute outlier detection. Second, they may not accurately locate outliers when multiple outliers exist in a cluster and correlate with each other. Third, the existing algorithms tend to abstract spatial objects as isolated points and do not consider their geometrical and topological properties, which may lead to inexact results.
This dissertation reports a study of the problem of abnormal spatial pattern recognition, and proposes a suite of novel algorithms. Contributions include: (1) formal denitions of various spatial outliers, including single-attribute outliers, multi-attribute outliers, and region outliers; (2) a set of algorithms for the accurate detection of single-attribute spatial outliers; (3) a systematic approach to identifying and tracking region outliers in continuous meteorological data sequences; (4) a novel Mahalanobis-distance-based algorithm to detect outliers with multiple attributes; (5) a set of graph-based algorithms to identify point outliers and region outliers; and (6) extensive analysis of experiments on several spatial data sets (e.g., West Nile virus data and NOAA meteorological data) to evaluate the eectiveness and eciency of the proposed algorithms.



 


 
 
 
 

You are cordially invited to attend Dai’s PhD Preliminary Exam defense
Oct. 30, 10am, 2006. Room 103



Speaker: Jing Dai
Advisor: Prof. Chang-Tien Lu


Title: Efficient Current Operations in Spaital Databases

Abstract:


Nowadays, demanded by the applications such as GIS, CAD, ecology analysis, and space research, efficient spatial data access methods have attracted a lot of research efforts. Especially, complex spatial operations and continuous spatial queries are becoming highlighted in spatial database area. However, most of the existing spatial query processing approaches were designed for single-user environments, which may not ensure the correctness and the data consistency in multiple-user environments. This research focuses on designing efficient concurrent operations on spatial data sets. Current multidimensional data access methods can be categorized into two types: 1) pure multidimensional indexing structure such as the R-tree family and grid file; 2) linear spatial access methods, represented by Space-Filling Curve (SFC) combined with B-trees. Concurrency control protocols have been designed for some pure multidimensional indexing structures, but none of them is suitable for variants of R-trees with object clipping, which are efficient in searching. On the other hand, there is no concurrency control protocol designed for linear spatial indexing structures, where the one-dimensional concurrency control protocols can not be directly applied.
In the proposed research, sound solutions for efficient concurrent access frameworks on both types of spatial indexing structures will be provided, as well as the spatial-temporal component, continuous query operations on moving objects, for multiple-user environment. These two concurrent access frameworks can satisfy the concurrency control requirements, meanwhile providing outstanding performance for concurrent queries. Major contributions include: (1) a new efficient spatial indexing approach with object clipping technique, ZR+-tree, that outperforms R-tree and R+-tree on searching; (2) complete concurrency control protocol, GLIP, to provide high throughput and phantom update protection on spatial indexing with object clipping; (3) efficient fundamental and complex concurrent operations for indexing based on linear spatial access methods, which form up the CLOCK protocol; (4) efficient concurrent continuous queries on moving objects for both concurrent spatial access framework.



 


 
 
 
 

You are cordially invited to attend Speer’s PhD Preliminary Exam defense
Oct. 20, 1pm, 2006. Room 320



Speaker: Anh Phan Speer
Advisor: Prof. Ing-Ray Chen


Title: Design and Analysis of Adaptive Fault Tolerant QoS Control Algorithms for Query Processing in Wireless Sensor Networks

Abstract:


Wireless sensor networks (WSNs) present several unique characteristics such as extremely resource-constrained sensors, large-scale random deployment, and data-centric communication protocols. These characteristics pose unprecedented challenges in the area of query processing in WSNs. This dissertation presents the design and validation of adaptive fault tolerant QoS control algorithms to achieve the desired quality of service (QoS) requirements and maximize the system lifetime in query-based WSNs.
Data sensing and retrieval in WSNs have a widespread application in areas such as security and surveillance monitoring, as well as command and control in battlefield situations. In query-based WSNs, a user would issue a query with QoS requirements in terms of reliability and timeliness, and expect a correct response to be returned within the deadline. Satisfying these QoS requirements requires that fault tolerance mechanisms through redundancy be used, which may cause the energy of the system to deplete quickly.
We analyze the effect of redundancy on the mean time to failure (MTTF) of a WSN with clustering, defined as the mean number of queries that the WSN is able to answer correctly until it fails due to channel faults, sensor faults, or sensor energy depletion. When the knowledge of query arrival rate is available, this metric can be translated into the conventional lifetime measure. In particular, we analyze the effect of redundancy on the MTTF of query-based cluster-structured WSNs. We show that a tradeoff exists between redundancy and MTTF. Furthermore, an optimal redundancy level exists such that the MTTF of the system is maximized.
We develop a hop-by-hop data delivery mechanism in which we utilize "source" and "path" redundancies with the goal to satisfy application QoS requirements while maximizing the lifetime of the WSN. When given QoS requirements of a query in terms of reliability and timeliness, we identify optimal "source" and "path" redundancies such that not only QoS requirements are satisfied, but also the lifetime of the system is prolonged. Numerical data are presented to demonstrate the feasibility of our approach.
To deal with network dynamics, we investigate proactive and reactive methods for cluster heads to dynamically collect channel and delay conditions to determine the optimal redundancy at runtime. We also design mechanisms to adapt to status changes of sensor nodes due to energy consumption and node failures. We validate our proposed adaptive fault tolerant QoS control algorithms with simulation studies based on J-Sim.



 


 
 
 
 

You are cordially invited to attend Vasudeo’s Master’s thesis defense
12:30pm,
May 5, 2006


Speaker: Mr. Jyotindra Vasudeo Vasudeo
Thesis Advisor: Prof. Greg Kulczycki


Title: The Design and Implementation of the Tako Language and Compiler

Abstract:


Aliasing complicates both formal and informal reasoning. Aliasing is a particular problem in object-oriented languages, where variables denote references to objects rather than object values. Researchers have proposed various approaches to the aliasing problem in object-oriented languages ranging from alias-control, where the effects of aliasing are isolated, to alias avoidance, where common sources of aliasing are avoided. The Tako language is an example of an object-oriented language that incorporates alias avoidance methodology. In this thesis we describe the design and implementation of the Tako language and compiler. We discuss the features of the language, describe the implementation of the Tako compiler in Java, and give proof rules for common statements in the language. We also present a brief case study that illustrates the paradigm shifts involved when moving from Java to Tako.



 


 
 
 
 

You are cordially invited to attend Salman’s Master’s thesis defense
10:00 AM, Monday,
May 23, 2005
 Room 113


Speaker: Mr. M. Salman Akram
salman at vt dot edu


Title: Managing Changes to Service Oriented Enterprises

Abstract:


In our thesis, we present a framework for managing changes in Service Oriented Enterprises (SOEs).  A service oriented enterprise outsources and composes its functionality from third-party Web service providers.  We focus on changes initiated or triggered by these member Web services. We present a taxonomy of changes that occur in service oriented enterprises.  We use a combination of several types of Petri nets to model the triggering changes and ensuing reactive changes.  The techniques presented in our work are implemented in WebBIS, a prototype for composing and managing e-business Web services.



 


 
 
 
 

You are cordially invited to attend the PhD thesis defense
1:00pm-3:00pm, Sep. 30th, 2005
 NVC Room 113


Speaker: Baoshan Gu
gubs@vt.edu


Title:
 Design and Analysis of Algorithms for Efficient Location and Service Management in Mobile Wireless Systems



Committee:


Dr. Ing-Ray Chen, Committee Chair
Dr. Luiz A. DaSilva
Dr. Denis Gracanin
Dr. Chang-Tien Lu
Dr. Scott F. Midkiff


Abstract:


Mobile wireless environments present several characteristics different from wired distributed systems, including unreliable wireless communication, disconnection, mobile and heterogeneous devices, and limited resources. These characteristics make design and validation of appropriate system supports for facilitating development of wireless mobile applications a challenging problem. This dissertation concerns two major system-support mechanisms in mobile wireless networks, namely, location management and service management.

Location management addresses the issues of how to track and locate a mobile user efficiently. Service management addresses the issues of how to efficiently deliver services to mobile user through limited wired and wireless network resources. This dissertation aims to design and analyze location and service management schemes that are efficient for cellular personal communication service (PCS) systems. We propose to address this research issue by considering three topics: location management, service management, and integrated location and service management.

A location management scheme must effectively and efficiently handle both user location-update and location-search operations. We propose to analyze existing location management algorithms by quantitatively evaluating the network signaling overhead for each of these algorithms and identifying conditions under which one algorithm may perform better than others. From insight leaned from the quantitative analysis, we design and analyze a hybrid algorithm that outperforms individual algorithms and show that such a hybrid algorithm can be uniformly applied to all mobile users with distinct call and mobility characteristics to simplify the system design without sacrificing performance.

For service management, we propose and analyze the notion of location-aware personal proxies with the goal to minimize the overall network signaling and communication cost caused by both location and service management operations. The idea is for each mobile user to use personal proxies as intelligent client-side agents to communication with services engaged by the mobile user. A personal proxy cooperates with the underlying location management system so that it is location-aware and can optimally decide when and how often it should move with the roaming user. We show that for cellular wireless networks that provide packet services, when given a set of model parameters characterizing the network and workload conditions, there exists an optimal proxy service area size for service handoffs such that the overall network signaling and communication cost for servicing location and service operations is minimized. These proxy-based mobile service management schemes are shown to outperform non-proxy-based schemes over a wide range of identified conditions. Moreover, when the mobile user is concurrently engaged in multiple services, the per-service proxy scheme, which uses a separate proxy for each service, outperforms the aggregate proxy scheme, which uses a single proxy to interact with multiple services taking their aggregate service characteristics into consideration.

Taking the lesson learned from designing location and service management algorithms, we investigate how location and service management can be more tightly integrated, i.e., by co-locating location databases with service proxies in order to further reduce the overall network signaling and communication cost due to managing both location and service operations. Four integrated location and service management schemes are proposed and analyzed in this dissertation for PCS cellular systems and validated with simulation and sensitivity analysis. We show that, when given an MH's mobility and service characteristics, there exists an optimal integrated location and service management scheme that would minimize the overall network communication cost as a result of executing the MH's location and service operations. We also demonstrate that the best integrated location and service scheme identified always performs better than the best decoupled scheme that considers location