A novel cache optimization algorithm and protocol for video streaming in pure peer-to-peer networks
Proceedings of the 2010 ACM workshop on Advanced video streaming techniques for peer-to-peer networks and social networking
| Hi-index | 0.00 |
In client-side segmented cache based P2P VOD system, statically cached video files in peers' disks can efficiently improve data acquisition rate. However, since disk space in a single peer is very limited, it becomes a key problem to design reasonable strategies for replacing those data segments. We start at studying the three VOD system models, and analyze the necessity of utilizing segmented storage mechanism, then focus on discussing and comparing five segment replacement strategies at client side: FIFO, LRU, LRF, EDF, and LSSF. These strategies sufficiently take local system parameters of a peer and global data distribution patterns into consideration, and manage to save high use rate, high hit rate segments in limited peers' disk space. Our experiment results indicate that, when a peer's cache size is increased from 240M to 960M, the use rate of data segments at client side rises by 36%, but the rate will rise little if continuously increasing its cache size; overall, under the same system environment, LSSF shows best performance