Optimizing Resource Scheduling in BitTorrent File Distribution Network
Zheng Chen, Jinyao Yan · 2009
In recent years, BitTorrent file distribution network has been more and more widely used for media file distribution. Its build-in resource scheduling policies (local rare first, tit-for-tat, etc.) work well in file distribution in single swarm environments. However, the resource scheduling policy is missing in multiple swarms environments and the resource utilization has not been optimized. In this work, our goal is to optimize the resource utilization in multiple swarms environments according to link utilization and peer satisfaction. We propose our optimizing resource policies based on swarm parameters, namely hidden-seed policy and weight-leecher policy. Our evaluation results show that the hidden-seed policy increases the link utilization while the weight-leecher policy improves the peer satisfaction in the system.