International Journal of Engineering and Technology

Torrent is used to download the same data by applying the bit torrent communication protocol. Some of the common terminologies used in torrent are peers, leechers, seeders and trackers. The peers are sometimes referred as leechers, the leecher is somebody who is currently downloading the file, the seeder is somebody who is currently uploading a file and tracker is the one who keeps track of all the activities of leecher and seeder. This paper mainly focuses on file sharing in cheap and effective peer to peer system, which suffers due to slow uploading and downloading and bandwidth lacking fairness. In this paper we report on experiments conducted to pin point the cause to increase the Bandwidth in uploading the data. And so we are proposing a new system using the k-nearest neighbour algorithm to ensure fairness. KNN algorithm is the simplest of all machines learning algorithms. By using this algorithm the objects are classified and grouped by neighbour’s voting providing rendering the shortest path among varying neighbours. K-is the smallest integer. If K=+1, then the object is grouped to the class of nearest neighbours. Thus our proposed methodology results in bandwidth fairness by increasing the performance of downloading and providing bandwidth in a uniform manner.