Abstract | ||
---|---|---|
Measuring and modeling peer to peer (P2P) traffic is of the key importance for the traffic engineering of P2P networks, because recent studies have revealed that the majority of the total traffic is the P2P traffic. Although numerous articles have been published in P2P traffic measurement, the mathematical modeling of the P2P traffic is an unexplored area. To address this, we propose a novel model called MAPP (Modified Alternating fractal renewal process for Peer to Peer) for P2P traffic. The model is a modification of the ON/OFF periods in the Alternating Fractal Renewal Process (AFRP). Our model can capture Long Range Dependence (LRD) and heavy tailedness. We show that the superposition of heterogenous MAPPs used for aggregated P2P traffic converges to a kind of alpha stable process. We provide the queuing analysis in both individual and superimposed MAPPs. The proposed model is evaluated with BitTorrent. Gnutella and eDonkey traffics. |
Year | Venue | Keywords |
---|---|---|
2009 | COMPUTER SYSTEMS SCIENCE AND ENGINEERING | Long Range Dependence,P2P Traffic,Alternating Fractal Renewal Process,Heavy Tail Distribution |
Field | DocType | Volume |
Renewal theory,Peer-to-peer,Computer science,Fractal,Artificial intelligence,Distributed computing | Journal | 24 |
Issue | ISSN | Citations |
6 | 0267-6192 | 0 |
PageRank | References | Authors |
0.34 | 0 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Kiarash Mizanian | 1 | 8 | 3.06 |
Mehdi Vasef | 2 | 6 | 2.41 |
Morteza Analoui | 3 | 124 | 24.94 |