Free riding on Gnutella

Adar, Eytan; Huberman, Bernardo A.

doi:10.5210/fm.v5i10.792

Cited by 1,153 publications

(813 citation statements)

References 7 publications

Supporting

Mentioning

783

Contrasting

Unclassified

Order By: Relevance

“…c) Cooperation: All peers do not share the same number of files and do not exhibit the same "social behavior". As observed in [23], a large proportion of the user population is made of so-called free-riders, who do not make any file accessible to other users and essentially behave as clients. On the other hand, a small proportion of the users (less than 5%) contribute more than two thirds of the files shared in the system and essentially behave as servers.…”

Section: B) Contentmentioning

confidence: 99%

“…On the other hand, a small proportion of the users (less than 5%) contribute more than two thirds of the files shared in the system and essentially behave as servers. Based on the study in [23], we assign the following storage capacity to the peers in the network: 70% of the peers do not share any file (free-riders); 20% share 100 files or less; 7% share between 101 and 1, 000 files; finally, 3% of the peers share between 1, 001 and 2, 000 files (actual storage capacities are chosen uniformly at random). With this distribution, we have observed in our experiments a total storage capacity of more than 1, 600, 000 files, with more than 150, 000 distinct files being shared.…”

Section: B) Contentmentioning

confidence: 99%

See 1 more Smart Citation

Efficient search in unstructured peer‐to‐peer networks

Cholvi

Felber

Biersack

2004

Trans Emerging Tel Tech

View full text Add to dashboard Cite

Abstract-The huge popularity of recent peer-to-peer (P2P) file sharing systems has been mainly driven by the scalability of their architectures and the flexibility of their search facilities. Such systems are usually designed as unstructured P2P networks, because they impose few constraints on topology and data placement and support highly versatile search mechanisms. A major limitation of unstructured P2P networks lies, however, in the inefficiency of their search algorithms, which are usually based on simple flooding schemes.In this paper, we propose novel mechanisms for improving search efficiency in unstructured P2P networks. Unlike other approaches, we do not rely on specialized search algorithms; instead, the peers perform local dynamic topology adaptations, based on the query traffic patterns, in order to spontaneously create communities of peers that share similar interests. The basic premise of such semantic communities is that file requests have a high probability of being fulfilled within the community they originate from, therefore increasing the search efficiency. We propose further extensions to balance the load among the peers and reduce the query traffic. Extensive simulations under realistic operating conditions substantiate that our techniques significantly improve the search efficiency and reduce the network load. I. INTRODUCTION A. MotivationsThe last few years have witnessed the appearance of a growing number of peer-to-peer (P2P) file sharing systems. Such systems make it possible to harness the resources of large populations of networked computers in a cost-effective manner, and are characterized by their high scalability.P2P file sharing systems mainly differ by their search facilities. The first hugely successful P2P data exchange system, Napster [1], incorporates a centralized search facility that keeps track of files and peer nodes; queries are executed by the central server, while the resource-demanding file transfers are performed using P2P communication. This hybrid architecture offers powerful and responsive query processing, while still scaling well to large peer populations. The central server needs, however, to be properly dimensioned to support the user query load. In addition, it constitutes a single point of failure and can easily be brought down in the face of a legal challenge, as was the case for Napster. Consequently, most recent P2P file sharing systems have adopted more decentralized architectures.Roughly speaking, the P2P networks that do not rely on a centralized directory can be classified as either structured or unstructured. Structured P2P networks (e.g., Chord [2], CAN [3],

show abstract

Section: B) Contentmentioning

confidence: 99%

Section: B) Contentmentioning

confidence: 99%

Efficient search in unstructured peer‐to‐peer networks

Cholvi

Felber

Biersack

2004

Trans Emerging Tel Tech

View full text Add to dashboard Cite

show abstract

“…Another type of locality characterizing peer-to-peer systems is the skewed level of service provided by different peers: a small number of peers respond to most queries, while a third to two thirds are "free riders" that do not contribute any service [14,311]. In particular, one study showed that the top responding peer could respond to nearly half of all queries.…”

Section: Peer-to-peer File Sharingmentioning

confidence: 99%

“…Examples include "weather" (14), "dictionary" (21), "lyrics" (64), and "horoscope" (86). Then there are quite a few nonsense queries, many of which are just parts of Internet addresses, such as "http" (13), ".com" (20), "www."…”

Section: Transactional Queriesmentioning

confidence: 99%

Workload Modeling for Computer Systems Performance Evaluation

Feitelson

2015

196

131

View full text Add to dashboard Cite

Reliable performance evaluations require the use of representative workloads. This is no easy task since modern computer systems and their workloads are complex, with many interrelated attributes and complicated structures. Experts often use sophisticated mathematics to analyze and describe workload models, making these models difficult for practitioners to grasp. This book aims to close this gap by emphasizing the intuition and the reasoning behind the definitions and derivations related to the workload models. It provides numerous examples from real production systems, with hundreds of graphs. Using this book, readers will be able to analyze collected workload data and clean it if necessary, derive statistical models that include skewed marginal distributions and correlations, and consider the need for generative models and feedback from the system. The descriptive statistics techniques covered are also useful for other domains.

show abstract

“…Incentives and micro-payments can be used to stimulate the users [1] and to avoid the free-riding phenomenon [2], and they are largely used in practice, e.g., BitTorrent [3], EMule [4] and Mojo-Nation 1 . In particular, a micro-payment scheme is an interesting alternative to a differential service incentives, especially when a market place is layered on top of a p2p system.…”

Section: Introductionmentioning

confidence: 99%

A P2P Market Place Based on Aggregate Signatures

Catalano¹,

Ruffo

Schifanella

2005

Parallel and Distributed Processing and Applications - ISPA 2005 Workshops

View full text Add to dashboard Cite

Abstract.A peer-to-peer market place is likely to be based on some underlying micro-payment scheme where each user can act both as a customer and as a merchant. Such systems, even when designed for largely distributed domains, may be implemented according to hybrid topologies where trusted third intermediaries (e.g. the broker) are single points of failures. For this reason it is crucial that such central entities scale well w.r.t. the overall number of transactions. In this paper, we focus on PPay as a case study, to show how the broker would greatly benefit in terms of computational cost if aggregate signatures are adopted instead of RSA signatures.

show abstract

Free riding on Gnutella

Cited by 1,153 publications

References 7 publications

Efficient search in unstructured peer‐to‐peer networks

Efficient search in unstructured peer‐to‐peer networks

Workload Modeling for Computer Systems Performance Evaluation

A P2P Market Place Based on Aggregate Signatures

Contact Info

Product

Resources

About