Large Scale Distributed Simulation of p2p Networks

Abstract

p2p systems have witnessed phenomenal development in recent years, evaluating and analysing new and existing new algorithms and techniques is a key issue for developers of p2p systems. In this context Simulation is an important tool for p2p developers. However, such systems are often very large and few existing simulators offer the ability to execute systems of real world size. In this paper we present a tool for executing large scale simulation of p2p systems which scale effectively, only limited by the amount of computational resource available (memory and CPU). This is achieved through the application of parallel discrete event simulation techniques to an existing, already scalable simulator, PeerSim. We show results from a case study using the Chord p2p protocol, indicating good scalability both in terms of size (memory) and execution time (CPU).

Performing accurate simulations for deadline-aware applications

“…However, their intrinsic set-up complexity undermines the simulator usability, particularly for inexperienced analysts, and its realistic models implementations might reduce significantly the scalability for cloud services, such as data transfer with rate control. According to a survey in Peer-to-Peer simulators by Dinh et al [4], the use of simulators like ns-3 is actually very occasional, and most analysts opt to develop their own tool. However, these tools are not necessarily made publicly available, becoming single paper or study-simulators, preventing simulation results reproductions.…”

Evaluating Large Scale Distributed Simulation of P2P Networks

“…We carried out a survey over the total of 76 papers on P2P system [3] 1 , the result of which indicates that simulation is the dominant analysis tool in structured P2P research.…”

“…Our distributed simulator dPeerSim [3] is based on PeerSim [9], an open source P2P simulator written in Java. It allows large scale simulation of P2P networks, providing two modes of operation: cycle driven and and event driven.…”

“…To reduce the overall number of null message, for each incoming channel on each LP, we remove redundant messages, so that only one null message (with highest timestamp) can exist at any given time [10]. More detailed description of the main extensions to PeerSim to enable distributed simulation can be found in [3].…”

“…In [3], we presented dPeerSim, a distributed simulation kernel for P2P networks based on PeerSim [9]. We demonstrated via preliminary experiments that using dPeerSim, scalability of tens of millions of nodes could be achieved, to the best of our knowledge for the first ever time [11].…”