No abstract

Modern web pages are no longer static plain HTML files, and they provide rich content through scripting languages such as JavaScript and multimedia. However, communications between browser applications that use scripting languages remain challenge. As a solution to this problem, we propose BISSA, a communication model which provides a unified and timedecoupled communication platform based on tuple spaces for browser applications. BISSA consists of an in-browser tuple space and a scalable and distributed peer-to-peer global tuple space, and both can either act standalone or collaborate with each other.The in-browser tuple space provides a solid communication infrastructure for web gadgets. The integration with the peerto-peer global tuple space further enforces this paradigm of communication, effectively allowing web gadgets to contribute or co-ordinate with the underlying computing infrastructure.This paper presents BISSA, its architecture and show how browser applications can use BISSA as an inter-gadget communication solution, storage platform for application generated data and as a middleware to develop web-based applications that brings the computation power of browser in to the grid.

No abstract

We focus on two classes of problems in graph mininghere: (1) finding trees and (2) anomaly detection using networkscan statistics in complex networks. These are fundamentalproblems in a broad class of applications. Most of the parallelalgorithms for such problems are either based on heuristics,which do not scale very well, or use techniques like colorcoding, which have a high memory overhead. In this paper, wedevelop a novel approach for parallelizing both these classesof problems, using an algebraic representation of subgraphsas monomials—this methodology involves detecting multilinearterms in multivariate polynomials. Our algorithms show goodscaling over a large regime, and they run on networks with closeto half a billion edges. The resulting parallel algorithm for treesis able to scale to subgraphs of size18, which has not beenshown before, and it significantly outperforms the best priorcolor coding based method (FASCIA) by more than two ordersof magnitude. Our algorithm for network scan statistics is thefirst such parallelization, and it is able to handle a broad class ofscan statistics functions (both parametric and non-parametric),with the same approach.

No abstract