This paper addresses the problem of churn---the continuous process of node arrival and departure---in distributed hash tables (DHTs). We argue that DHTs should perform lookups quickly and consistently under churn rates at least as high as those observed in deployed P2P systems such as Kazaa. We then show through experiments on an emulated network that current DHT implementations cannot handle such churn rates. Next, we identify and explore three factors affecting DHT performance under churn: reactive versus periodic failure recovery, message timeout calculation, and proximity neighbor selection. We work in the context of a mature DHT implementation called Bamboo, using the ModelNet network emulator, which models in-network queuing, cross-traffic, and packet loss. These factors are typically missing in earlier simulationbased DHT studies, and we show that careful attention to them in Bamboo's design allows it to function effectively at churn rates at or higher than that observed in P2P file-sharing applications, while using lower maintenance bandwidth than other DHT implementations.

Of the many P2P file-sharing prototypes in existence, BitTorrent is one of the few that has managed to attract millions of users. BitTorrent relies on other (global) components for file search, employs a moderator system to ensure the integrity of file data, and uses a bartering technique for downloading in order to prevent users from freeriding. In this paper we present a measurement study of BitTorrent in which we focus on four issues, viz. availability, integrity, flashcrowd handling, and download performance. The purpose of this paper is to aid in the understanding of a real P2P system that apparently has the right mechanisms to attract a large user community, to provide measurement data that may be useful in modeling P2P systems, and to identify design issues in such systems.

Despite its popularity, relatively little is known about the traffic characteristics of the Skype VoIP system and how they differ from other P2P systems. We describe an experimental study of Skype VoIP traffic conducted over a five month period, where over 82 million datapoints were collected regarding the population of online clients, the number of supernodes, and their traffic characteristics. This data was collected from September 1, 2005 to January 14, 2006. Experiments on this data were done in a black-box manner, i.e., without knowing the internals or specifics of the Skype system or messages, as Skype encrypts all user traffic and signaling traffic payloads. The results indicate that although the structure of the Skype system appears to be similar to other P2P systems, particularly KaZaA, there are several significant differences in traffic. The number of active clients shows diurnal and work-week behavior, correlating with normal working hours regardless of geography. The population of supernodes in the system tends to be relatively stable; thus node churn, a significant concern in other systems, seems less problematic in Skype. The typical bandwidth load on a supernode is relatively low, even if the supernode is relaying VoIP traffic.

During recent years, peer-to-peer (P2P) file-sharing systems have evolved in many ways to accommodate growing numbers of participating peers. In particular, new features have changed the properties of the unstructured overlay topology formed by these peers. Despite their importance, little is known about the characteristics of these topologies and their dynamics in modern file-sharing applications. This paper presents a detailed characterization of P2P overlay topologies and their dynamics, focusing on the modern Gnutella network. Using our fast and accurate P2P crawler, we capture a complete snapshot of the Gnutella network with more than one million peers in just a few minutes. Leveraging more than 18,000 recent overlay snapshots, we characterize the graph-related properties of individual overlay snapshots and overlay dynamics across hundreds of back-to-back snapshots. We show how inaccuracy in snapshots can lead to erroneous conclusions—such as a power-law degree distribution. Our results reveal that while the Gnutella network has dramatically grown and changed in many ways, it still exhibits the clustering and short path lengths of a small world network. Furthermore, its overlay topology is highly resilient to random peer departure and even systematic attacks. More interestingly, overlay dynamics lead to an “onion-like ” biased connectivity among peers where each peer is more likely connected to peers with higher uptime. Therefore, long-lived peers form a stable core that ensures reachability among peers despite overlay dynamics. 1

Recently, peer-to-peer (P2P) networks have emerged as an attractive solution to enable large-scale content distribution without requiring major infrastructure investments. While such P2P solutions appear highly beneficial for content providers and end-users, there seems to be a growing concern among Internet Service Providers (ISPs) that now need to support the distribution cost. In this work, we explore the potential impact of future P2P file delivery mechanisms as seen from three different perspectives: i) the content provider, ii) the ISPs, and iii) individual content consumers. Using a diverse set of measurements including BitTorrent tracker logs and payload packet traces collected at the edge of a 20,000 user access network, we quantify the impact of peer-assisted file delivery on end-user experience and resource consumption. We further compare it with the performance expected from traditional distribution mechanisms based on large server farms and Content Distribution Networks (CDNs).

Abstract not found

Peer-to-Peer networks are gaining increasing attention from both the scientific and the large Internet user community. Popular applications utilizing this new technology offer many attractive features to a growing number of users. At the heart of such networks lies the search algorithm. Proposed methods either depend on the network-disastrous flooding and its variations or utilize various indices too expensive to maintain. In this paper, we describe an adaptive, bandwidth-efficient algorithm for search in unstructured Peer-to-Peer networks, the Adaptive Probabilistic Search method (APS). Our scheme utilizes feedback from previous searches to probabilistically guide future ones. It performs efficient object discovery while inducing zero overhead over dynamic network operations. Extensive simulation results show that APS achieves high success rates, increased number of discovered objects, very low bandwidth consumption and adaptation to changing topologies. 1.

We consider the problem of choosing who to "befriend " among a collection of known peers in distributed P2P systems. In particular, our work explores a number of P2P protocols that, by considering peers' lifespan distribution a key attribute, can yield systems with performance characteristics more resilient to the natural instability of their environments.

Performance of peer-to-peer resource sharing networks depends upon the level of cooperation of the participants. To date, cash-based systems have seemed too complex, while lighter-weight credit mechanisms have not provided strong incentives. We propose exchange-based mechanisms for providing incentives for cooperation in peer-to-peer file sharing networks. Peers give higher service priority to requests from peers that can provide a simultaneous and symmetric service in return. We generalize this approach to #-way exchanges among rings of peers and present a search algorithm for locating such rings. We have used simulation to analyze the effect of exchanges on performance. Our results show that exchange-based mechanisms can provide strong incentives for sharing, offering significant improvements in service times for sharing users compared to free-riders, without the problems and complexity of cash- or credit-based systems.

Desktop resources are attractive for running computeintensive distributed applications. Several systems that aggregate these resources in desktop grids have been developed. While these systems have been successfully used for many high throughput applications there has been little insight into the detailed temporal structure of CPU availability of desktop grid resources. Yet, this structure is critical to characterize the utility of desktop grid platforms for both task parallel and even data parallel applications. We address the following questions: (i) What are the temporal characteristics of desktop CPU availability in an enterprise setting? (ii) How do these characteristics affect the utility of desktop grids? (iii) Based on these characteristics, can we construct a model of server "equivalents" for the desktop grids, which can be used to predict application performance ? We present measurements of an enterprise desktop grid with over 220 hosts running the Entropia commercial desktop grid software. We utilize these measurements to characterize CPU availability and develop a performance model for desktop grid applications for various task granularities, showing that there is an optimal task size. We then use a cluster equivalence metric to quantify the utility of the desktop grid relative to that of a dedicated cluster.