Distributed Data Streams Indexing using Content-based Routing Paradigm

Report ID: 
2004-01
Authors: 
Ahmet Bulut, Roman Vitenberg and Ambuj K. Singh
Date: 
2004-01-01 04:00:00

Abstract

We are witnessing a dramatic increase in the use of data-centric distributedsystems such as global grid infrastructures, sensor networks, networkmonitoring, and various publish-subscribe systems. The visions of massivedemand-driven data dissemination, intensive processing, and intelligentfusion in order to build dynamic knowledge bases that seemed infeasible justa few years ago are about to come true. However, the realization of thispotential demands adequate support from middleware that could be used todeploy and support such systems.We propose an integrated distributed indexing architecture that supportsscalable handling of intense dynamic information flows. The suggestedindexing scheme is geared towards providing timely responses to queries ofdifferent types with varying precision requirements while minimizing the useof network and computational resources. Our solution bestows thecapabilities provided by content-based routing schemes, such as scalabilityand load balancing of communication as well as adaptivity in presence ofdynamic changes. The paper elaborates on database and content-based routingmethodologies used in the integrated solution as well as non-trivialinteraction between them, thereby providing a valuable feedback to thedesigners of these techniques. We analyze the scalability of the proposedarchitecture by running simulation tests with our prototype implementationon top of the Chord system simulator.

Document

2004-01.pdf