Modeling Interactions in a P 2 P Network

15th EUROMICRO International Conference on Parallel, Distributed and Network-Based Processing (PDP'07), 2007

Peer-to-peer (P2P) networks can reduce the distribution cost of large media files for the original provider of the data significantly. Thereby, the BitTorrent protocol is widely used in the Internet today. Most research work studies the protocol analytically, by simulations at the flow-level or real world experiments. Thereby, for flow-level simulations the influence of neglecting packet-level characteristics is not yet quantified. Therefore, this paper compares packet-level simulation results with flow-level values and analytically derived bounds. Our findings show that BitTorrent is near to optimal at flow-level for different scenarios. Naturally, packet-level results deviate more from the optimal values but differences are at most around 30% in our simulations. Furthermore, we show that the propagation delay can significantly influence the download performance of BitTorrent.

2006

Today, a large part of the traffic carried by the Internet is generated by file-sharing applications based on the P2P paradigm. Moreover, P2P traffic has particular features, quite different from the ones generated by "classical" data applications, and generally causes highly dynamic and distributed traffic matrixes that heavily stress network infrastructures. With this respect, it is reasonable to say that the analysis and study of the overall network performance have to take into account P2P traffic characteristics and behavior. To this purpose, the present paper approaches P2P traffic modeling activities, by describing a Gnutella simulator developed at the Telecommunication Networks and Telematics (TNT) Laboratory of DIST. The aim is well founded on the partial lack of simulators that effectively and completely describe the quantity, the characteristics and the impact of P2P traffic in networks.

2002

Despite recent excitement generated by the peer-to-peer (P2P) paradigm and the surprisingly rapid deployment of some P2P applications, there are few quantitative evaluations of P2P systems behavior. The open architecture, achieved scale, and self-organizing structure of the Gnutella network make it an interesting P2P architecture to study. Like most other P2P applications, Gnutella builds, at the application level, a virtual network with its own routing mechanisms. The topology of this virtual network and the routing mechanisms used have a significant influence on application properties such as performance, reliability, and scalability. We have built a "crawler" to extract the topology of Gnutella's application level network. In this paper we analyze the topology graph and evaluate generated network traffic. Our two major findings are that: (1) although Gnutella is not a pure power-law network, its current configuration has the benefits and drawbacks of a power-law structure, and (2) the Gnutella virtual network topology does not match well the underlying Internet topology, hence leading to ineffective use of the physical networking infrastructure. These findings guide us to propose changes to the Gnutella protocol and implementations that may bring significant performance and scalability improvements. We believe that our findings as well as our measurement and analysis techniques have broad applicability to P2P systems and provide unique insights into P2P system design tradeoffs.

2010

The peer-to-peer (P2P) paradigm has greatly influenced the design of Internet applications nowadays. It gained both user popularity and significant attention from the research community, aiming to address various issues arising from the decentralized, autonomous, and the self-organizing nature of P2P systems [379]. In this regard, quantitative and qualitative analysis at large scale is a crucial part of that research. When evaluating widely deployed peer-to-peer systems an analytical approach becomes, however, ineffective due to the large number of simplifications required. Therefore, conclusions about the real-world performance of P2P systems can only be drawn by either launching an Internet-based prototype or by creating a simulation environment that accurately captures the major characteristics of the heterogeneous Internet, e.g. round-trip times, packet loss, and jitter.

European Transactions on Telecommunications, 2004

Peer-to-peer (P2P) services evolved to the most popular applications in today's Internet. In particular, P2P networks became very popular amid the relentless spread of Gnutella, Kazaa, eDonkey and BitTorrent file sharing applications. Remarkably, only very simple protocols and almost no support by the transport network was required to make these distributed services operable on a large scale in very little time. P2P entails a highly attractive paradigm in distributed computing: P2P is based on communication between equals. The peers are highly autonomous. P2P services provide simple and efficient mechanisms to pool and share exchangeable resources like CPU cycles, disk space or content. These features facilitate that any peer can be removed without resulting in a complete loss of service, in contrast to the traditional client/server concept, where a failure of the central control entity may corrupt the service completely. A further reason for the notable success of P2P is that these networks operate on the application level and typically form application-specific overlays. P2P overlays work without particular network or transport support, and can be run completely at the edge of a network. They apply self-organizing principles and special routing schemes to locate and exchange resources. While P2P overlays do implement a certain type of group communication structure, they do not suffer from the same deployment difficulties as multicast services did in the past. The P2P mode of operation however, also has some downsides. P2P protocols cause higher traffic volumes, including data traffic as well as signaling traffic, since the peers need to synchronize. P2P network topologies reveal high variability and traffic patterns of P2P applications fluctuate considerably in time and space. Traditional network design techniques and traffic engineering procedures have to be validated to work in the P2P context before they are applied. New engineering methods may be required in order to maintain the autonomous and self-organizing characteristics of P2P and to provide appropriate service stability, quality and efficiency for P2P services. The aim of this special issue of the European Transactions on Telecommunication is to present very recent, outstanding research work on P2P networks and P2P services. The Call for Paper resulted in the submission of 42 papers from Europe, the United States and the Asian-Pacific region. The editor team has selected eight of these submissions after a thorough review process and extensive discussions for publication in this special issue. The guest editor team of this ETT special issue would also like to thank the publisher, the reviewers and the authors for their efforts and flexibility to produce, review and finalize this special issue in a very short time frame. We hope that this special issue may serve to promote further research in this new and popular area of telecommunication. The contribution of Azzouna and Guillemin evaluates the impact of P2P applications on traffic in wide area networks. The authors analyze measurements taken on a high speed IP backbone link carrying P2P applications traffic towards several ADSL areas. They observe that the prevalent part, the traffic is caused by P2P applications (almost 80% of total traffic) and that the usage of network becomes highly symmetric. Since long lasting TCP connections of P2P applications have a rather small bit rate and do not show evidence for long range dependency, the global traffic can be described by means of usual teletraffic models based on M=G=1 queues with Weibullian service times. The paper of Schollmeier and Schollmeier investigates P2P traffic by a graph theoretical study of the properties of P2P overlays. Analytical topology models for infinite and finite random networks are used to derive performance characteristics for P2P protocols and to evaluate the effect of introducing centralized entities, such as superpeers. Cholvi, Felber and Biersack propose novel mechanisms for improving the search efficiency in unstructured P2P networks. The mechanism is based on peers performing local dynamic topology adaptations, which are derived from query traffic patterns. The topology adaptation creates spontaneously semantic communities, i.e. groups of peers that share similar interests, in which file requests are fulfilled more efficiently. Copyright # 2004 AEI Wierzbicki et al. investigate the possibility of caching of files in P2P file sharing applications. They show that the file popularity in P2P file sharing networks does not follow Zipf's law and that P2P file sharing traffic differs strongly from well-studied web traffic. The results are used to investigate the efficiency of different cache management policies for the FastTrack P2P file sharing protocol. Sasabe et al. investigate scalable media search and in-time retrieval methods based on a P2P network architecture. The method is able to achieve continuous media play-out for popular media streams without introducing extra load on the system. The authors also investigate the failure of LRU cache replacement algorithms for continuous media play-out of unpopular media streams and propose a biological-inspired cache replacement mechanism that considers the balance between supply and demand for media streams. The algorithm can adapt easily to changes in the popularity of media streams. Kumar et al. present a P2P network based on a butterfly overlay topology. The diameter of this overlay achieves a theoretical lower bound of log n=log n log n, if the average routing table size is not more than log n. Compared to existing DHT schemes, the diameter is reduced by a factor of log log n. This reduction translates into the same amount of reduction on query latency and average traffic per node. The authors show that the proposed scheme maintains the same level of robustness as existing DHT schemes. Yamada, Hoshiai and Kimura propose brokerless trust and reputation mechanisms for the COMNet P2P framework. COMNet is an intelligent community spaces architecture, which supports the activation of communities through P2P technology. The proposed trust and reputation algorithms are solely based on information from other peers. Bisignano, Di Modica and Tomarchio present a framework for mobile P2P applications, which combines P2P ad hoc communication with the capabilities of small mobile units. The framework is designed for J2ME enabled mobile devices in an infrastructure-less environment. One of the key features of the framework is the interoperability with JXTA, a wellknown P2P open platform. A simple but effective example of application is provided showing the viability of the designed framework.