An intelligent peer-to-peer streaming model for minimizing flash crown

Abstract

Peer-to-peer streaming systems (P2PSS) have become well deployed over the Internet in recent times due to its robustness, scalability, distributive nature and cost effectiveness. It is not un-common that peers arrive on the system in large numbers at a particular time in order to retrieve multimedia files. The large spike in the number of peers arriving at a time can be referred to as flash crowd. Several methods have been proposed, models have been designed aiming at providing solution to the problem. This paper attempts to alleviate flash crowd that may occur in the system using a fuzzy logic control system. Inputs were created and fuzzified, rules were developed and then the outputs were defuzzified. Analysis of the results derived from the MATLAB simulation reveal that under a flash crowd scenario, our fuzzy logic controller functions appropriately by the detecting flash crowd when it is about to occur and then necessary actions are taken. Also, with the fuzzy logic control system peer access in and out of the system is successfully controlled. The system allows for new peers to connect more with high bandwidth peers aiming at making the upload rate of each parent proportional to its upload bandwidth thereby utilizing the limited bandwidth resources of the entire P2P system more effectively.

Keywords: Peer-to-peer networks, video streaming and flash crowd

Vol. 26, No. 1, June, 2019