Research Article
Extremal Search of Decision Policies for Scalable Distributed Applications
@INPROCEEDINGS{10.4108/infoscale.2007.207, author={Gang Chen and Chor Ping Low and Zhonghua Yang}, title={Extremal Search of Decision Policies for Scalable Distributed Applications}, proceedings={2nd International ICST Conference on Scalable Information Systems}, proceedings_a={INFOSCALE}, year={2010}, month={5}, keywords={}, doi={10.4108/infoscale.2007.207} }- Gang Chen
Chor Ping Low
Zhonghua Yang
Year: 2010
Extremal Search of Decision Policies for Scalable Distributed Applications
INFOSCALE
ICST
DOI: 10.4108/infoscale.2007.207
Abstract
The ongoing trend of constructing open, scalable distributed systems such as peer-to-peer (P2P) systems demands for ef- fective tools to manage the interactions between constituent entities (or nodes). One such tool is through imposing de- cision policies at the network level. However very few tech- niques are available to allow computers autonomously iden- tify good policies with limited human intervention. In this paper, we propose an Extremal Programming (EP) algo- rithm to achieve automatic policy identification. The algo- rithm is inspired by recent advances in understanding far from equilibrium phenomena in terms of self-organized crit- icality (SOC). The effectiveness of EP is evaluated through a P2P application called location-aware video streaming (LAVS). The simulation studies in LAVS demonstrate that with EP, the fast and effective sharing of video streams is achieved.