Design and Implementation of a Data Stream Management System

Malempati, N (2012) Design and Implementation of a Data Stream Management System. Masters thesis, Indian Institute of Technology, Hyderabad.

Preview

Text
CS10M03.pdf - Submitted Version
Download (2MB)

Abstract

The amount of data stored by companies has grown exponentially over the last decade. Of late, data is being continuously collected for various purposes - click stream analysis, credit card transactions for fraud detection, weather monitoring, stock tickers in nancial services, link statistics in networking, user logins and web surng statistics, highway trac congestion analysis and so on. The data that is being collected is in the form of a stream - arrives continuously, at a variable rate, and can occupy potentially innite storage. As organizations have realized that fast and ecient processing of this data can help in protable predictions, there exists a need for developing systems to handle this collected data eectively. We present in this thesis, the architecture of a generic Database Stream Management System (DBSMS) to handle streaming data. While literature has provided insights into Data Stream Management Systems (DSMS), the DBSMS is a dierent approach that tries to integrate a DSMS with the traditional Database Management Systems (DBMS). We discuss the need for such a generic DBSMS and present the system that we have implemented using the discussed architecture. We also present the performance of our system, in terms of space taken, time taken to answer a query and the accuracy of the result compared to a DBMS. Finally, we conclude with brief discussion on certain goals and open challenges that are of interest and which still need to be addressed by the system.

[error in script]

IITH Creators: