DOCSLIB.ORG

Fairtorrent: Bringing Fairness to Peer-To-Peer Systems

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Fairtorrent: Bringing Fairness to Peer-To-Peer Systems FairTorrent: Bringing Fairness to Peer-to-Peer Systems Alex Sherman, Jason Nieh, and Cliff Stein Department of Computer Science Columbia University Technical Report CUCS-029-08 May, 2008 {asherman, nieh, cliff}@cs.columbia.edu Abstract to take as much as possible from the system while con- tributing little resources, and a lack of quality-of-service The lack of fair bandwidth allocation in Peer-to-Peer guarantees to support streaming applications. systems causes many performance problems, includ- Fair bandwidth allocation in P2P systems can be diffi- ing users being disincentivized from contributing upload cult to achieve for several reasons. First, bandwidth re- bandwidth, free riders taking as much from the system sources belong to and are controlled by individual peers, as possible while contributing as little as possible, and a not by a single party. Unlike a router or a server, there lack of quality-of-service guarantees to support stream- is no central entity that controls and arbitrates access ing applications. We present FairTorrent, a simple dis- to all resources. Second, bandwidth resources are dis- tributed scheduling algorithm for Peer-to-Peer systems tributed all over the Internet, and therefore vary geo- that fosters fair bandwidth allocation among peers. For graphically and also are limited by the various Internet each peer, FairTorrent maintains a deficit counter which Service Providers. Third, the amount of bandwidth re- represents the number of bytes uploaded to a peer mi- sources available is not known in advance and peers can- nus the number of bytes downloaded from it. It then not be relied upon to specify their own resources hon- uploads to the peer with the lowest deficit counter. Fair- estly. Fourth, bandwidth resources may vary over time Torrent automatically adjusts to variations in bandwidth for several reasons, including network conditions, mo- among peers and is resilient to exploitation by free-riding bile peers connecting at different access points, and a peers. We have implemented FairTorrent inside a BitTor- user using the available bandwidth for other activities. rent client without modifications to the BitTorrent proto- Finally, any fair allocation mechanism must be strong col, and compared its performance on PlanetLab against enough to withstand attempts by free riders to manipu- other widely-used BitTorrent clients. Our results show late the system. that FairTorrent can provide up to two orders of magni- tude better fairness and up to five times better download Of course, fairness is not the only desirable property performance for high contributing peers. It thereby gives in a peer-to-peer system, users also desire good perfor- users an incentive to contribute more bandwidth, and im- mance, measured by download time. We show that im- prove overall system performance. proving fairness leads to improved performance, some- times, with substantial gains. 1 Introduction Many approaches have attempted to address the prob- In the past decade, the usage of Peer-to-Peer (P2P) file- lem of fair bandwidth exchange. One approach, which is sharing applications on the Internet has experienced ex- employed by the popular file-sharing system BitTorrent, plosive growth. Many users and businesses now rely on is to use a tit-for-tat (TFT) heuristic. However, recent P2P file-sharing for distributing videos, software, and studies [27, 21, 15] have demonstrated weaknesses in the documents. Although P2P file-sharing is now an inte- BitTorrent TFT mechanism that can be exploited by free- gral part of our overall computing experience, these ap- riding clients to take advantage of high-bandwidth con- plications are plagued by a fundamental problem of un- tributors. For example, the LargeView exploit client [27] fairness in how bandwidth among peers is used and al- takes advantage of the fact that BitTorrent uses a fraction located. Unfairness causes many performance problems, of its bandwidth to upload optimistically to randomly including users being disincentivized from contributing chosen peers. By opening connections to many peers, the upload bandwidth, a growing number of free riders, users LargeView client allows a large number of its neighbors who cap their upload bandwidth to zero or a small value to pick it as the target for their optimistically-spent band- 1 width. As another example, the BitTyrant client [21] is Tyrant [21], a modified Azureus 2.5 client that was opti- based on the observation that even though BitTorrent at- mized for improved download performance. tempts to find reciprocating neighbors, it is often willing For peers with widely different bandwidths across to upload to peers who reciprocate at a much lower rate. a uniform distribution, our measurements demonstrate BitTyrant attempts to find peers where it can maximize that FairTorrent provides more than an order of mag- the difference between what it receives from them versus nitude better fairness and up to 50% faster download what it uploads to them. performance compared to other BitTorrent implementa- To address these problems, we present FairTorrent, tions. Furthermore, FairTorrent makes more efficient use a new P2P scheduling algorithm that fosters fair band- of available upload bandwidth and provides download width allocation among peers. In its simplest form, at times that are strongly correlated with the peers’ upload each peer, FairTorrent maintains a deficit counter for each capacities, incentivizing peers to upload more. For a high neighboring peer which represents the number of bytes it bandwidth uploader in the presence of many low con- uploaded to this neighbor minus the number of bytes it tributors, our measurements demonstrate that FairTorrent downloaded from it. When it is ready to send a packet can provide two orders of magnitude better fairness and of data, FairTorrent identifies the peer with the lowest up to five times faster download performance compared deficit counter and sends data to that peer. to other BitTorrent implementations. We also show that Our main result is that this surprisingly simple ap- FairTorrent provides improved fairness and performance proach provides much better fairness while simultane- even if low contributors run other BitTorrent clients, in- ously decreasing most download times, often by a sig- cluding clients designed to exploit unmodified BitTorrent nificant amount. Download times are also more corre- for better performance. lated with the peers’ upload rates, thereby incentivizing This paper describes the design, implementation, and peers to upload more. FairTorrent has a number of ad- evaluation of FairTorrent. Section 2 discusses related ditional useful properties: (1) FairTorrent provides fair work. Section 3 presents FairTorrent, which is analyzed bandwidth allocation, operating only at individual peers, in Section 4. Section 5 presents experimental results on in a distributed manner that does not require any central- PlanetLab comparing FairTorrent performance to other ized control of peers or other P2P resources. (2) Fair- BitTorrent implementations. Finally, we present some Torrent does not need to measure available download concluding remarks. bandwidth, allocate precise upload or download rates for any peers, or rely on estimates or advanced knowledge 2 Background and Related Work of available bandwidth from users or other peers. (3) BitTorrent [7] employs a rate-based tit-for-tat (TFT) FairTorrent allows a peer to automatically adjust to dy- heuristic to incentivize peers to upload and attempts to namic network conditions and any changes in the rate at provide fair exchange of bandwidth between peers. Peers which peers contribute bandwidth, avoiding long discov- participating in the download of the same target file form ery times of like-uploading peers as evidenced in BitTor- a swarm. The target file is conceptually broken up into rent [21]. (4) FairTorrent provides low variance in band- pieces, typically 256 KB in size. Peers tell one another width allocation, enabling it to be useful for streaming which pieces of the target file they already have and re- applications. (5) FairTorrent has no magic parameters quest missing pieces from one another. Requests are typ- and requires no tuning, is simple to implement, and re- ically made for 16 KB sub-pieces. Peers that already quires no changes to the BitTorrent protocol, making it have the entire file are seeds. Peers that are still down- easy to deploy with existing P2P systems. loading pieces of the file are leechers. TFT is used in We have analyzed, implemented, and measured Fair- a swarm to enable fair bandwidth exchange during the Torrent to evaluate its effectiveness. Our analysis results current download of a file. It operates by having each show that FairTorrent runs efficiently, using O(log k) BitTorrent client upload to N other peers in round-robin time per transmission in a network with k peers. We fashion, where N − k of the peers have provided the best have proved for a small network that the algorithm has download rate during the most recent time period, and k fairness bounds independent of the amount of data sent, peers are randomly selected to help discover other peers and conjecture that a similar result holds for any network. with similar upload rates. (N is typically between 5 and We have implemented FairTorrent inside a BitTorrent 10). The set of peers to which a client uploads is periodi- client without changing the BitTorrent protocol, making cally changed based on measurements of their download it compatible with existing BitTorrent implementations. rates. BitTorrent refers to the selection and deselection of We evaluated its performance on PlanetLab against three a peer for uploading as unchoking and choking, respec- other BitTorrent implementations, the original BitTorrent tively. Python client by Bram Cohen [7], the latest version of Much work has been done in studying the behavior of the popular Azureus Java BitTorrent client [3], and Bit- BitTorrent. Qiu and Srikant [23] show that under some 2 bandwidth distributions, the system eventually converges download, one model proposes to optimize average per- to a Nash equilibrium.
Load more

Recommended publications
  • Replication Strategies for Streaming Media
    “replication-strategies” — 2007/4/24 — 10:56 — page 1 — #1 Research Report No. 2007:03 Replication Strategies for Streaming Media David Erman Department of Telecommunication Systems, School of Engineering, Blekinge Institute of Technology, S–371 79 Karlskrona, Sweden “replication-strategies” — 2007/4/24 — 10:56 — page 2 — #2 °c 2007 by David Erman. All rights reserved. Blekinge Institute of Technology Research Report No. 2007:03 ISSN 1103-1581 Published 2007. Printed by Kaserntryckeriet AB. Karlskrona 2007, Sweden. This publication was typeset using LATEX. “replication-strategies” — 2007/4/24 — 10:56 — page i — #3 Abstract Large-scale, real-time multimedia distribution over the Internet has been the subject of research for a substantial amount of time. A large number of mechanisms, policies, methods and schemes have been proposed for media coding, scheduling and distribution. Internet Protocol (IP) multicast was expected to be the primary transport mechanism for this, though it was never deployed to the expected extent. Recent developments in overlay networks has reactualized the research on multicast, with the consequence that many of the previous mechanisms and schemes are being re-evaluated. This report provides a brief overview of several important techniques for media broad- casting and stream merging, as well as a discussion of traditional IP multicast and overlay multicast. Additionally, we present a proposal for a new distribution system, based on the broadcast and stream merging algorithms in the BitTorrent distribution and repli- cation system. “replication-strategies” — 2007/4/24 — 10:56 — page ii — #4 ii “replication-strategies” — 2007/4/24 — 10:56 — page iii — #5 CONTENTS Contents 1 Introduction 1 1.1 Motivation .
    [Show full text]
  • Bittorrent Files Hace Stopped Downloading
    bittorrent files hace stopped downloading Why Some Torrents Don’t Download? A torrent that doesn’t start downloading or one that suddenly stops can be very frustrating. You check your Internet connection, the cables, and everything looks good. So what can be the reasons for those torrents that don’t seem to work? Some Torrents Don’t Download. The main reason behind a torrent file that doesn’t even start downloading is the lack of seeders and peers. In other words, there is no one seeding that file, meaning there’s no place where you can download it from. That’s why it’s very important that you have a look at the number of seeders and peers every time you start a new download. Seeders are the users who already finished downloading and are only sharing. The peers are the ones like you, the ones who are downloading and uploading at the same time. Some Files Suddenly Stop Downloading. This is another common scenario. We’ve all been there when a torrent stops at some moment, such as 99%. That usually happens when there are only peers, but no seeders . If you think about it, it makes total sense. The peers have many parts of the torrent in common, and they will share those between them. But because there are zero seeders, no one has the entire file , and everyone will share the same parts and stop in the same percentage point. A Dead Torrent. Both of the situations we just saw are what users in the community call a “dead torrent”.
    [Show full text]
  • Improving Fairness in Peer-To-Peer Networks by Separating the Role of Seeders in Network Infrastructures
    Turkish Journal of Electrical Engineering & Computer Sciences Turk J Elec Eng & Comp Sci (2016) 24: 2255 { 2266 http://journals.tubitak.gov.tr/elektrik/ ⃝c TUB¨ ITAK_ Research Article doi:10.3906/elk-1402-304 Improving fairness in peer-to-peer networks by separating the role of seeders in network infrastructures Alireza NAGHIZADEH∗, Reza EBRAHIMI ATANI Department of Computer Engineering, Faculty of Engineering, University of Guilan, Rasht, Iran Received: 27.02.2014 • Accepted/Published Online: 08.07.2014 • Final Version: 15.04.2016 Abstract:Fairness is one of the most important challenges that should be considered as a priority when designing a P2P file-sharing network. An unfair P2P network may attract free-riders, frustrate the majority of users, and consequently shorten the longevity of files. When BitTorrent protocol was first introduced, by suggesting new algorithms like tit-for- tat or rarest-first and combining them with methods like choking and unchoking, it pushed fairness one step forward. However, BitTorrent did not bring a plenary solution and has its own shortcomings. For instance, neither this protocol nor other P2P protocols that we are aware of precisely indicate how seeders should be treated in their networks. Most of the time they dictate that seeders upload as much as possible. With this approach, seeders can easily be abused by free-riders. It gets even worse when we realize that for lack of a good infrastructure a lawful user may become a free-rider unknowingly. The purpose of this paper is to address this problem by suggesting a novel and universal method that can be implemented in current P2P networks.
    [Show full text]
  • P2P File Sharing P2P File Sharing
    P2P File Sharing P2P file sharing Alice chooses one of the peers, Bob. File is copied from Bob’s PC to Example Alice’s notebook: HTTP Alice runs P2P client While Alice downloads, other application on her notebook users uploading from Alice. computer Alice’s peer is both a Web client Intermittently connects to and a transient Web server. Internet; gets new IP address All peers are servers = highly for each connection scalable! Asks for “Hey Jude” Application displays other peers that have copy of Hey Jude. P2P: centralized directory (Napster’s Approach) Bob centralized directory server original “Napster” design 1 1) when peer connects, it peers 1 informs central server: IP address 1 3 content 2 1 2) Alice queries for “Hey Jude” 3) Alice requests file from Alice Bob P2P: problems with centralized directory file transfer is Single point of failure decentralized, but locating content is Performance highly decentralized bottleneck Copyright infringement Query flooding: Gnutella overlay network: graph edge between peer X fully distributed and Y if there’s a TCP no central server connection public domain all active peers and protocol edges is overlay net many Gnutella clients Edge is not a physical implementing protocol link Given peer will typically be connected with < 10 overlay neighbors Gnutella: protocol Ì File transfer: Query message HTTP sent over existing TCP connections Query Ì peers forward QueryHit Query message Ì QueryHit sent over reverse Query path QueryHit Scalability: limited scope flooding Gnutella: Peer joining 1. Joining peer X must find some other peer in Gnutella network: use list of candidate peers 2.
    [Show full text]
  • Compsci 514: Computer Networks Lecture 21-2: from Bittorrent to Bittyrant Problem Statement
    CompSci 514: Computer Networks Lecture 21-2: From BitTorrent to BitTyrant Problem Statement ... Server • One-to-many content distribution – Millions of clients downloading from the same server Evolving Solutions • Observation: duplicate copies of data are sent • Solutions – IP multicast – End system multicast – Content distribution networks e.g. Akamai – P2P cooperative content distribution • Bittorrent etc. IP multicast • End systems join a multicast group • Routers set up a multicast tree • Packets are duplicated and forwarded to multiple next hops at routers • Multicast pros and cons End system multicast • End systems rather than routers organize into a tree, forward and duplicate packets • Pros and cons Content distribution networks • Akamai – Works well but expensive, requires infrastructure support Peer-to-Peer Cooperative Content Distribution u Use the client’s uplink bandwidth u New problem: incentives for cooperation or how to motivate clients to upload The Gnutella approach u All nodes are true peers u A peer is the publisher, the uploader and the downloader. u No single point of failure. u Efficiency and scalability issue: u File searches span across a large number of nodes generating lots of traffic. u Integrity, i.e.content pollution issue: u Anyone can claim that he publishes valid content u No guarantee of quality of objects u Incentive issue: u No incentives for cooperation (free-riding in Gnutella) Outline u Problem of Content Distribution u The BitTorrent approach u BitTyrant BitTorrent overview Tracker 1 2 3 Leecher A Seeder Leecher C Leecher B u File is divided into chunks (e.g. 256KB) uShA1 hashes of all the pieces are included in the .torrent file for integrity check uA chunk is divided into sub-pieces to improve efficiency u Seeders have all chunks of the file u Leechers have some or no chunks of the file BitTorrent overview Tracker 1 2 3 Leecher A Seeder Leecher C Leecher B u File is divided into chunks (e.g.
    [Show full text]
  • Exploiting Bittorrent for Fun (But Not Profit)
    Exploiting BitTorrent For Fun (But Not Profit) Nikitas Liogkas, Robert Nelson, Eddie Kohler, and Lixia Zhang University of California, Los Angeles fnikitas, rlnelson, kohler, [email protected] ABSTRACT the file can be downloaded from different peers. A meta- This paper assesses BitTorrent’s robustness against selfish peers, data file is associated with every download. This file con- who try to download more than their fair share by abusing existing tains information necessary for the download process, in- protocol mechanisms. We design and implement three selfish-peer cluding the number of pieces and hashes for all the pieces; exploits and evaluate their effectiveness on public and private tor- the hashes are used by peers to verify that a piece has been rents. In practice, BitTorrent appears quite robust against this kind received correctly. This metadata file is typically created by of exploit: selfish peers can sometimes obtain more bandwidth, and the content provider, who must also launch at least one client honest peers’ download rates suffer slightly in consequence, but we offering the entire file for download. In order to join the observe no considerable degradation of the system’s quality of ser- download process, a client retrieves the metadata file out of vice. We identify private-torrent scenarios in which a selfish peer band, usually either from a well-known website or by email. could benefit more significantly at the expense of honest peers, and It then contacts the tracker, a centralized component that discuss the BitTorrent protocol mechanisms that lead to robustness by rendering these scenarios infeasible. keeps track of all the peers participating in the download.
    [Show full text]
  • Peer-To-Peer Networking
    Peer-to-Peer Networking Case Study: BitTorrent Lecture Content BitTorrent (protocol version 1.0) File transfer protocol Novel techniques for distributing content P2P system studied further The original design Later enchantments are not discussed here BitTorrent (BT) BT is a file transfer protocol for content distribution Protocol specifications v1.0 studied here A centralised P2P system (compare to Napster) A tracker server managing users© downloads BitTorrent concentrates on efficient file transfer Searching content is not provided by the protocol specification, but out-of-band methods Addresses the free riding problem in P2P file sharing BT Terminology (1/3) Torrent Set of peers cooperating to download the same content using BT protocol Tracker Centralised server keeping track of current participants. Does not involve to data transfers, but collect statistics Pieces and blocks File is cut into fixed size pieces (typically 256 KB) and pieces are further cut into blocks (transfer unit, 16 KB) BT Terminology (2/3) Metainfo file, or .torrent file Contains information about the file, its length, name and the address and port of the tracker Hashes for pieces of files for verification Interested and Choked A is marked as interested in peer B when B has pieces that A wants, and vice versa. Also, A is choked, when B decides not to upload data when A is interested. When B is willing to upload again, A is unchoked. BT Terminology (3/3) Peer set or a swarm The group of machines that are collectively connected for a particular file, i.e. a list of open TCP connections Leecher and seed Leecher is a peer that is still downloading the pieces.
    [Show full text]
  • Peer-To-Peer Networks – Protocols, Cooperation and Competition
    Peer-to-Peer Networks – Protocols, Cooperation and Competition Hyunggon Park Signal Processing Laboratory (LTS4), Institute of Electrical Engineering, Swiss Federal Institute of Technology (EPFL), Lausanne, Switzerland Rafit Izhak Ratzin Computer Science Department, University of California, Los Angeles (UCLA), Los Angeles, USA Mihaela van der Schaar Multimedia Communications and Systems Laboratory, Electrical Engineering Department, University of California, Los Angeles (UCLA), Los Angeles, USA 1. INTRODUCTION Peer-to-peer (P2P) networks connect many end-hosts (also referred to as peers) in an ad-hoc manner. P2P networks have been typically used for file sharing applications, which enable peers to share digitized content such as general documents, audio, video, electronic books, etc. Recently, more advanced applications such as real-time conferences, online gaming, and media streaming have also been deployed over such networks. Unlike traditional client-server networks, where servers only provide content, and clients only consume content, in P2P networks, each peer is both a client and a server. It has been observed that P2P file sharing applications dominate Internet traffic usage. In fact, a wide range of measurements, which were performed in 8 different geographic regions during the years of 2008-2009, show that P2P networks generated most of the traffic in all monitored regions, ranging from 43% in Northern Africa to 70% in Eastern Europe (http://www.ipoque.com/). The same study also identified that BitTorrent (Cohen, 2003) is the most popular protocol on the Internet, generating most of the traffic in 7 out of 8 regions ranging from 32% in South Africa to 57% in Eastern Europe. The details of the BitTorrent protocol will be discussed in Section 3.
    [Show full text]
  • Libswift P2p Protocol: an Analysis and Extension
    Master Thesis - TRITA-ICT-EX-2012-262 LIBSWIFT P2P PROTOCOL: AN ANALYSIS AND EXTENSION Fu Tang Design and Implementation of ICT products and systems Royal Institute of Technology (KTH) [email protected] October 30th, 2012 Supervisor: Flutra Osmani Examiner: Bj¨ornKnutsson Department: ICT School, NSLab Royal Institute of Technology (KTH) 1 Abstract More and more end-users are using P2P protocols for content sharing, on-demand and live streaming, contributing considerably to overall Internet traffic. A novel P2P streaming protocol named libswift was developed to enable people experience a better service by consuming less resources and transferring less unnecessary functions and metadata. This master thesis studies the inner functioning of libswift and analyzes some of the vulnerabilities that directly impact performance of the protocol, namely download speed and response delay. By investigating the behavior of libswift in scenarios with multiple peers, we found that the lack of a peer selection mechanism inside the protocol affects download efficiency and response time. We also discovered that libswift's internal piece picking algorithm raises competition among peers, thus not fully utilizing connected peers. In addition, we found that current libswift implementation does not follow the specification for PEX peer discovery, thus we modified PEX algorithm to support another message that is used to proactively request new peers from the currently connected. Having made these observations, we designed and implemented a peer selection extension interface that allows for third-party peer selection mechanisms to be used with libswift protocol. Apropos, we tested the interface (or adapter) with an example peer selection mechanism that groups peers according to properties such as latency and locality.
    [Show full text]
  • Velocizzare Emule Abc.Txt” È Una Stronzata (+ Raccolta Di Guide Serie Alla Configurazione E Al Corretto Uso Di Emule E a Come Ottimizzarne Le Prestazioni)
    Ecco perchè il contenuto del file “velocizzare emule abc.txt” è una stronzata (+ raccolta di guide serie alla configurazione e al corretto uso di eMule e a come ottimizzarne le prestazioni) Qualche persona dall'intelligenza discutibile ha diffuso su emule un file dal nome “velocizzare emule ABC.txt”. Il contenuto di questo files è un misto di stronzate e di cose parzialente vere ma se si seguono alla lettera le istruzioni si peggiorano le prestazioni di eMule e si fanno una serie di azioni dannose per la rete stessa che vengono considerate spam. Questo è il contenuto del sudetto file: ci sono solo 7 semplici passi per velocizzare emule il più possibile. 1- innanzi tutto dovete avere parecchi file condivisi, per fare questo andate in opzioni > cartelle e selezionate le cartelle che volete condividere ad esempio quelle che cintengono la musica. per avere una buona velocità dovreste avere almeno 500 file condivisi, si avete capito bene, 500. 2- ora che i vostri file condivisi sono tanti, selezionateli tutti e con il tasto destro scegliete priorità > *release*, così facendo darete la possibilità agli altri di scaricare più velocemente e il vostro punteggio aumenterà permettendovi di connettervi più velocemente ai server, e di scalare le code 3- togliete i limiti di download e di upload in opzioni > connessione, prima di fare questo assicuratevi di fare il wizard della vostra connessione cliccando su wizard e selezionate il vostro tipo di connessione 4- in opzioni > opzioni avanzate, attivate trova automaticamente limite di upload in upload dinamico e , più in basso, allungate la vostra coda a 6000 punti.
    [Show full text]
  • Architecture and Download Behavior of Xunlei:A Measurement-Based Study
    Architecture and Download Behavior of Xunlei:A Measurement-Based Study Min ZHANG Wolfgang JOHN Changjia CHEN Beijing Jiaotong University Chalmers University of Technology Beijing Jiaotong University Beijing, China, 100044 Goteborg, Sweden Beijing, China, 100044 [email protected] [email protected] [email protected] Abstract—this paper presents a measurement-based study of companies such as Google, Geyuan Ventures and IDGVC. Xunlei, a young but increasingly popular P2P application Although Xunlei does not provide an official English currently mainly used in China. Due to its proprietary version (yet) and only runs on Windows machines, it gained protocols and data encryption, packet-level data was analyzed significant popularity in regions outside China as well, in order to infer the network architecture and the superior mainly due to Chinese migrants [1]. Furthermore, translated download behavior, Xunlei's main key to success. We identify and ported versions are likely to be available soon, Xunlei's core-server layer with a clear division of tasks. We eventually allowing Xunlei to spread worldwide. found that Xunlei increases download speeds by applying Besides regional limitations, we consider that Xunlei's multi-source technology in an effective way, which includes proprietary nature and use of data encryption are the main (inappropriate) direct linking. Finally, we reveal that Xunlei utilizes UDP as main transport protocol both for bulk file cause for the very limited attention of the scientific transfer and IPTV streaming. As the first comprehensive community. In this paper, we seek to fill this gap by study on Xunlei, this paper provides new insights for continuing an earlier study [3] in order to provide a more researchers and application designers, and contributes to a detailed study on Xunlei.
    [Show full text]
  • Free Riding in Bittorrent Is Cheap
    Free Riding in BitTorrent is Cheap Thomas Locher1, Patrick Moor2, Stefan Schmid1, Roger Wattenhofer1 1 Computer Engineering and Networks Laboratory (TIK), ETH Zurich, 8092 Zurich, Switzerland {lochert, schmiste, wattenhofer}@tik.ee.ethz.ch 2 Google Inc., Mountain View, CA 94043, USA [email protected] ABSTRACT while seeders (“altruistic peers”) clearly offer the oppor- tunity to freeload, we are even able to download content While it is well-known that BitTorrent is vulnerable to self- quickly if we ignore seeders and download solely from ish behavior, this paper demonstrates that even entire files other peers that do not possess all pieces of the desired can be downloaded without reciprocating at all in BitTor- content (leechers). This implies that the basic piece ex- rent. To this end, we present BitThief, a free riding client change mechanism does not effectively restrain peers from that never contributes any real data. First, we show that freeloading. simple tricks suffice in order to achieve high download Sharing communities are also investigated in this paper. rates, even in the absence of seeders. We also illustrate By banning users with constantly low sharing ratios or by how peers in a swarm react to various sophisticated at- denying them access to the newest torrents available, such tacks. Moreover, our analysis reveals that sharing com- communities encourage users to upload more than they munities—communities originally intended to offer down- download, i.e., to keep their sharing ratio above 1. We will loads of good quality and to promote cooperation among show that sharing communities are particularly appealing peers—provide many incentives to cheat.
    [Show full text]