Full metadata record
|dc.contributor||Department of Electronic and Information Engineering||en_US|
|dc.publisher||Hong Kong Polytechnic University||-|
|dc.rights||All rights reserved||en_US|
|dc.title||Modeling, analyzing and improving the performance of BitTorrent swarming systems||en_US|
|dcterms.abstract||BitTorrent is one of the most popular peer-to-peer content distribution systems, and plays a dominant role with respect to the Internet traffic. Although BitTorrent is very effective in terms of bandwidth utilization, it is confronted with a serious problem that in many BitTorrent swarms, peers cannot complete the download due to the lack of some content blocks. Therefore, it is very important to find solutions to this problem, which we call content availability, as they can significantly enhance the service capability and performance of BitTorrent swarms. This work aims to develop an insightful understanding to the performance of BitTorrent swarming systems, and explore how it can be improved, with a special focus on content availability. In this study, we first perform a comprehensive study on the modeling and analysis of BitTorrent swarms. We derive the closed-form expressions for the performance metrics of Bit-Torrent swarms related to content availability, and investigate the influence of bundling on content availability. It is shown that bundling could greatly improve the availability of content, and that in a bundled swarm, peers could complete the download earlier than they would do in the individual swarm, given an appropriate number of files are bundled. In addition, the altruistic behavior of peers is also studied. We present an analysis on how peers' altruistic behavior affects the length of the residual active period after the leave of the publisher, and quantify the impact of bundling on the residual active period in the presence of peers' altruistic behavior.||en_US|
|dcterms.abstract||Next, we carry out an in-depth investigation on the feasibility of using network coding to ameliorate content availability of BitTorrent swarms. We first present a mathematical analysis on the potential improvement in the content availability and bandwidth utilization induced by two existing network coding schemes. The analysis reveals that network coding has a large potential to improve content availability, but both of the existing two schemes are not feasible as they either incur a very high coding complexity and disk operation overhead or cannot effectively leverage the potential of improving content availability. In this regard, a simple sparse network coding scheme is proposed, which addresses both the drawbacks in the existing schemes, and a new block scheduling algorithm is also developed in order to accommodate the proposed coding scheme into BitTorrent. The extensive simulation results demonstrate the effectiveness of the proposed coding scheme in terms of improving content availability. Finally, as motivated by the recent development of private BitTorrent communities, we conduct a detailed survey on one of the largest private BitTorrent communities, CHDBits. First, we characterize torrents from the perspectives of age, size, popularity and average user download rate, and then profile the different aspects of CHDBits users, e.g., diurnal access pattern, user traffic, seeding and leeching time. We also develop an in-depth understanding to how CHDBits users participate in downloading and uploading. The survey results suggest some new findings with regard to user behavior: low bandwidth users are more likely to participate in torrents with a smaller content size or a higher popularity, and compared with low bandwidth users, high bandwidth users tend to participate in more torrents, but spend less time in seeding.||en_US|
|dcterms.extent||xviii, 150 leaves : ill. ; 30 cm.||en_US|
|dcterms.isPartOf||PolyU Electronic Theses||en_US|
|dcterms.LCSH||Peer-to-peer architecture (Computer networks)||en_US|
|dcterms.LCSH||Downloading of data.||en_US|
|dcterms.LCSH||Hong Kong Polytechnic University -- Dissertations||en_US|
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item: