Completed MS Projects/2007-2008
From CSWiki
[edit] Florencio A. Almirol: CoolStateLA Enterprise Systems Analysis and Architecture
Advisor: Dr. Russell Abbott
CoolStateLA is a new vision to develop a CSULA converged newsroom that will consist mainly of the University Times Newspaper, CoolStateLA News Website, CoolStateLA News Web Radio, and CoolStateLA News Webcast. The focus of this analysis is on the process of producing the news website from the CoolStateLA Enterprise Information System.
The spiral software development process was used to approach the project. Requirements were collected in Use Cases and the functionality was analyzed. Components were designed in an architectural diagram. The conceptual model was designed through a database schema. Artifacts were iteratively refined by communicating with all of the stakeholders – News Director, technical adviser, and development team members.
The results have been a better understanding of the news industry and Fedora. Baseline prototypes were developed for CoolStateLA.com, the asset manager, and workflow manager. Communication and iteration are the keys to analyzing and designing the complex converged newsroom that provides information to change lives.
The Presentation, Architectural Diagram, and Database Schema on CoolStateLA Enterprise Systems Analysis and Architecture.
[edit] Chao Lin Chu: Spider Search: an Efficient and Non-Frontier-Based Real-Time Search Algorithm
Advisor: Dr. Russell Abbott
Real-time search algorithms are limited to constant-bounded search at each time step. We do not see much difference between standard search algorithms and good real-time search algorithms when problem sizes are small. However, having a good real-time search algorithm becomes important when problem sizes are large. There are some well known real-time search algorithms, for instance, Learning Real-time A* (LRTA*). But LRTA* is inadequate for problems that require deep path analysis at each time step. In general, any frontier-based search will be inadequate when deep path analysis is required. In this paper we introduce a simple yet efficient algorithm, Spider Search, which uses very low constant time and space to solve problems when agents need deep (but not exhaustive) path analysis at each step. We expect that Spider search is the first in a new class of tree-based rather than frontier-based search algorithms.
The Presentation on Spider Search: an Efficient and Non-Frontier-Based Real-Time Search Algorithm.
[edit] Joan Hyo Kim: Unintended consequences
Advisor: Dr. Russell Abbott
Several distinct applications of unintended consequence have been presented in articles and books over the last several decades. To date, there remains a level of inconsistency in its application. As a result unintended consequence lacks a clear, distinctive characterization. A general notion of unintended consequence could simply be an unplanned outcome. The goal of this thesis is to survey preceding concepts of unintended consequences and to propose a distinction between unintended consequences arising as a result of installed mechanisms and those which occur as historical contingencies; the contingent sequence of events that occur. We propose an unintended consequence occurs when an unexpected and unintended use is made of a mechanism or formalism—such as a law, a rule, a regulation, a computer program or even a custom or an accepted ethical or moral precept—that has been established in the world. Presumably the mechanism or formalism was established with the intent of achieving some particular positive result or reducing or eliminating something that is considered negative. The unintended consequence is almost always different from that end and in many cases may be contrary to it. Lastly, the exploit creates a definite advantage or profit.
The Presentation on Unintended Conseqeuences.
[edit] Mohammed Saleh: Pair-wise key establishment for wireless sensor network
Advisor: H. Guo
Managing security measures is a critical requirement, particularly in the application of sensor networks. In order to improve security in wireless sensor networks (WSN) it is necessary to encrypt messages before transmission. Point-to-point communication uses and end-to-end cryptography that achieves a high level of security. This is accomplished by distributing keys among all end points. The issue then becomes incompatibility with passive participation of the local broadcast. Close interaction of the nodes, short distances between them, energy constrains, and the unattended deployment of the system make WSN highly vulnerable to node capture. This paper attempts to establish a new pair-wise key schema that addresses the initialization process. The new model is loosely based on an existing model known as SHELL. We are proposing a pair-wise key establishment of neighboring sensor nodes that will improve the efficiency of the network architecture and reduces vulnerability to attacks and compromises. Our schema can be effectively applied to provide protection for the communication of neighboring nodes in the network; sensor nodes are capable of providing the same level of security as before while reducing the consequences of node compromise.
The presentation on Pair-wise key establishment for wireless sensor network.
[edit] Farrukh Shakil: Asset Management System For CoolStateLA
Advisor: Dr. Russell Abbott
The Contents in any Converged Newsroom are not just simple chunk of data. They are some kind of Intellectual Work in the form of Digital Assets. That Intellectual Work requires a mechanism to be recognized and effectively communicated to others. A Converged Newsroom like CoolStateLA not only needs to handle many different media types like images, audio, video and text but also needs to represent Intellectual Work in the form of Complex Media Objects composed of either other Complex Media Objects or simple medias. Such a system would require a very efficient and powerful Asset Manager. This Thesis proposes an Asset Management System that supports all the necessary tasks and decisions surrounding ingesting, defining, cataloging, organizing, storage and retrieval of CoolStateLA's digital contents in the form of Intellectual Work. It suggests an Object-Oriented Model at Data tier and the Controller built as a part of this thesis to be used for Business Logic. In addition to that this Thesis proposes the Content Models that allow the use of Web Services to effectively communicate the Intellectual Work to others.
The Presentation on Asset Management System for CoolStateLA.
[edit] Chanwit Suebsureekul: Instant Messaging for CSNS
Advisor: C. Sun
Computer Science Network Services (CSNS) is a web application that provides services for students, faculty, and staff in the Department of Computer Science at California State University, Los Angeles. This project adds an additional service to CSNS. It provides a means of communication among CSNS users via a real-time communication scheme, Instant Messaging (IM). The communication is done based on text messages transmitted via computer network such as the Internet.
Instant Messaging for CSNS offers many features that will encourage the communication among members of the CS department. A user can start a conversation with other users provided that he knows the usernames (screen names) of the other parties. The conversation could be between two people (private talk) or more (group discussion). Since usernames are required for communicating with other users, IM for CSNS provides a buddy list (contact list) where each user can store and manage a list of buddies' screen names. From the buddy list, the user can see each buddy's online status. The online status tells the user whether the buddy is available to talk or not. All of the above functionalities are presented to users through an enhanced, user-friendly user interface. Generally speaking, IM for CSNS is a web application that supports all common features found in traditional IM applications.
Although it is a web application, the system employs Ajax and JavaScript Drag & Drop API to give the look and feel of the conventional GUI application to users. Besides page design issues, the design of server side architecture is important as well. Unlike traditional web application, this project is a Comet style application where web server can push data without clients explicitly request for it. This thesis discusses about a Comet implementation called Pushlet and how it can be used in the application to provide IM services to users.
Finally, the thesis discusses methods used for testing the application. Testing is done in two aspects including feature testing and performance testing. In feature testing, a test plan containing usage scenarios was created and used for testing various web browsers to see if they support all IM features provided by the application. In performance testing, a JMeter test plan was created and used to simulate user behaviors to find out how many users the system can support.
The presentation on Instant Messaging for CSNS, March 13, 2008.
[edit] Sumathie Sundaresan: Survey of Privacy Protection for Medical Data
Advisor: H. Guo
In this paper, we will analyze seven different methodologies that are in practice for privacy protection of medical data. First method is Generalization, which generalizes the value of the quasi-identifier attributes. Second method is k-anonymity, where a table provides k-anonymity if the information for each person contained in the release cannot be distinguished from at least k-1 individuals. Third method is l-diversity, where a table is l-diverse if each equivalence class of the table has l-diversity with respect to the sensitive attribute. Fourth method is t-closeness, where a table is said to be t-closeness if the distance between the distributions of the attribute in the whole table is no more than a threshold t. Fifth method is m-invariance, which supports re publication of microdata after it has been updated with insertions and deletions. Sixth method is Personalized Privacy Preservation, which performs the minimum generalization for satisfying everybody’s requirements, and thus, retains the largest amount of information from the microdata. The last method is Anatomy, which releases all the Quasi-Identifiers and sensitive values in to two separate tables, which gives the accurate result for range queries.
The presentation on Survey of Privacy Protection for Medical Data
[edit] Yvonne Yao: An Improved Indexing Scheme for Range Queries
Advisor: H. Guo
In the Database-as-a-Service Model, sensitive data often raise the concern of leakage. Thus numerous techniques have been proposed to protect such data. Many of these techniques involve storing the encrypted representation of the data onto the database. Privacy-preserving can be achieved by this approach. However, querying the data over the encrypted representation becomes difficult. In this paper, we have proposed an indexing scheme, based on Query Optimal Bucketization, for range queries over encrypted numerical data. Our scheme, Deviation Bucketization, uses two levels of indices, with the first level being the partitions of the actual values, and the second level being the partitions of the deviation of the values in the first to the average of each partition. With two levels of buckets, we can further partition the values in each bucket. Since each bucket is further divided, less number of false positives will be returned. By comparing with Query Optimal Bucketization, the size of the superset produced by Deviation Bucketization is much smaller. The client will then have fewer records to decrypt, and hence, the performance is improved. We have implemented such scheme, and analyzed its performance against that of Query Optimal Bucketization. Experimental results have justified our claim.
The presentation on An Improved Indexing Scheme for Range Queries
