Bittorrent in a P2P social network

Bittorrent in a P2P social network

• 1.
  Dr.-Ing. Kalman Graffi Email:[email protected] Fachgruppe Theorie verteilter Systeme Fakultät für Elektrotechnik, Informatik und Mathematik Universität Paderborn Fürstenallee 11, D-33102 Paderborn, Deutschland Tel.+49 5251 606730, Fax. +49 5251 606697 http://www.cs.uni-paderborn.de/fachgebiete/fg-ti.html 4. Juli 2013 Sharing large files through Bit Torrent in Peer- Peer social networks Kailaash Balachandran Project Group - A Peer-to-Peer Framework for Social Networks Kailaash Balachandran [email protected]
• 2.
  PG – APeer-to-Peer Framework for Social Networks 2 Overview Written by Bram Cohen in 2001. Designed to transfer large files over the network. More than 160 million bit torrent clients with 100 million active users. Installed on 28.20% of computers worldwide in 2007. Hundreds of Index sites. Torrentz has alone indexed has 11.70 million torrents with 33.58 million connected peers. The average download speed is about 30 KB/s. Responsible for 25-35% of all internet traffic and 78% of P2P traffic.
• 3.
  PG – APeer-to-Peer Framework for Social Networks 3 Client – Server Downloading The browser requests the server for a file. The server sends back the requested file. Download speed affected by a number of factors such as  amount of traffic on the server the number of computers downloading the file. the demands of the server. Client Server Handling multiple requests
• 4.
  PG – APeer-to-Peer Framework for Social Networks 4 What is Bit-Torrent ? Bit-Torrent is a peer to peer (P2P) protocol used for sharing files over the network. A P2P client is used on every participating nodes. No central resource allocation. A single data is accessed effectively by hundreds of peers. The data is chopped into pieces and further into sub pieces. Every peer acts as both, client and server. One of Bit Torrent's most powerful idea is the choking mechanism.
• 5.
  PG – APeer-to-Peer Framework for Social Networks 5 Bit-Torrent Terminologies (1) Peer Active node in the network. Seeder Peer that has the complete file. Leecher Peer that still downloads pieces of the content. Swarm Group of peers active in the network. Includes both seeder and leecher.
• 6.
  PG – APeer-to-Peer Framework for Social Networks 6 Bit-Torrent Terminologies (2) Tracker A server that keeps track of leechers and seeds in the network. The tracker is centralized. .Torrent file : A small meta info file that has information about the content. Doesn‘t contain the content itself. Index sites : Also referred to as a torrent search engine. Contains .torrent files for download.
• 7.
  PG – APeer-to-Peer Framework for Social Networks 7 The Working
• 8.
  PG – APeer-to-Peer Framework for Social Networks 8 Piece Selection Stratergy (1) Strict Priority A piece is divided into many sub pieces. When a sub piece is requested, all remaining sub pieces are requested ASAP. Rarest First The peer selects the piece which is found in rare in its connected network. Makes sure that all pieces are available on the network.
• 9.
  PG – APeer-to-Peer Framework for Social Networks 9 Piece Selection Stratergy (2) Random piece first Initial leecher has got nothing to upload. Finding rarest piece may take long time. The peer selects a random piece to download. End Game Mode The peer sends requests to all connected piece for the missing piece. Piece gets downloaded from peer that has higher transfer rates. Cancel messages are sent to all appending peers.
• 10.
  PG – APeer-to-Peer Framework for Social Networks 10 Piece Selection Strategy (3) Piece selection strategy of a peer from download start to end.
• 11.
  PG – APeer-to-Peer Framework for Social Networks 11 Peer Selection Strategy (1) Every peer is responsible to maximize its download rate. Selecting a peer with good transfer rates is critical. The Choking Algorithm Choking = Refusal to Upload A peer A chokes peer B if it decides not to upload to B but download from B continues to happen. Follows Tit for tat mechanism. Prevents Free riding peers.
• 12.
  PG – APeer-to-Peer Framework for Social Networks 12 Peer Selection Strategy (2) A peer unchokes a number of connected peers (4 peers) to exchange pieces. 3 peers with fastest upload rate and 1 peer chosen randomly (optimistic unchoke). A peer purely based on their transfer rate. Calculated for every 10 seconds. Optimistic Unchoking A random peer is unchoked at every 30 seconds.
• 13.
  PG – APeer-to-Peer Framework for Social Networks 13 Peer Selection Strategy (3) Why ? Regular unchoking method may miss out some unused connections. New peers with good transfer rates are more likely to appear. Allows the system to evaluate the download capacity of new peers. Anti Snubbing A peer will get choked for longer duration. Such peer is assumed to be snubbed. Only a optimistic unchoke can release the peer from being snubbed.
• 14.
  PG – APeer-to-Peer Framework for Social Networks 14 Challenges Faced The Availability problem. System Integrity. Free riding still exists. Bit-Torrent is not completely de-centralized. Decentralized search. Improve social collaboration.
• 15.
  PG – APeer-to-Peer Framework for Social Networks 15 Bit-Torrent in LifeSocial (1) Guaranteeing the availability of the file Less than 4 percent of the peers have an uptime of over 10 hours. Solution = Provide more incentives such as social recognition and awards. Maintaining System Integrity All peers cannot be trusted. Solution = Achieve trust among peers. Friend to Friend network (F2F) paradigm.
• 16.
  PG – APeer-to-Peer Framework for Social Networks 16 Bit-Torrent in LifeSocial (2) Social trust becomes P2P trust. In F2F, connections are only possible among friends. Treat Before Trick The file to be shared is encrypted with a key k. Each piece of the file has a sub key. These sub keys have no dependence on the piece they have been sent. Initial seeder uploads both pieces and sub keys.
• 17.
  PG – APeer-to-Peer Framework for Social Networks 17 Bit-Torrent in LifeSocial (3) Leechers follows the rule ‘ Data First Key Later’ for bartering. Every peer in the network to exchange data to receive the key. Decentralized Trackers Modern Bit Torrent clients can exchange their neighbourhood sets through Peer Exchange (PEX). Track peers in the swarm by RePex(Reconnect to peers encountered before). A Repex peer includes swarm peers in the response. .
• 18.
  PG – APeer-to-Peer Framework for Social Networks 18 Bit-Torrent in LifeSocial (4) The 2-Hop TorrentSmell Algorithm. Hop 1 :
• 19.
  PG – APeer-to-Peer Framework for Social Networks 19 Bit-Torrent in LifeSocial (5) Hop 2 :
• 20.
  PG – APeer-to-Peer Framework for Social Networks 20 Bit-Torrent in LifeSocial (6) BuddyCast Algorithm Uses Gossip Protocol. A user’s file sharing history is shared with its connected peers. By exchanging these histories with others, the user builds a collection of a few hundred or more of such histories. If they are similar, they become taste buddies. Peers that become taste buddies can recommend files to each other.
• 21.
  PG – APeer-to-Peer Framework for Social Networks 21 Bit-Torrent in LifeSocial (7) Collaborative Downloads Each peer has to take one of two roles; either a Collector or a helper. The collector selects the best source from the helpers thus optimizing its download performance.
• 22.
  PG – APeer-to-Peer Framework for Social Networks 22 Conclusion Bit-torrent is a great technology for file sharing in a P2P network. A paradigm of bit torrent implementation has been presented. Listed a few challenges with possible solutions in context to LifeSocial. Future work.