Lazy caching: A novel proxy caching algorithm for peer-to-peer live streaming

Peer-to-Peer (P2P) live streaming systems are becoming popular and are imposing a heavy burden on Internet Services Providers (ISPs). Proxy caching has been shown to be an effective means of reducing operation costs for ISPs. Although many caching algorithms for conventional web applications and P2P file sharing systems have been proposed and deployed, there have been few works on the caching for P2P live streaming. Data requests in P2P live streaming have distinct characteristics; for example, they are concentrated in a limited window which progresses rapidly with time in a monotonic mode. In this paper, we propose a novel caching algorithm for P2P live streaming called Lazy Caching (LC), which is designed based on the characteristics of P2P live streaming. In LC, data requests that encounter cache misses are stored in a Request Buffer (RB) that is processed periodically. LC can better utilize the cache space than traditional caching algorithms at the cost of introducing a response latency which is at most one RB processing period. Experimental results show that the hit ratio of LC, at a processing period of 0.5 second, can be as much as 20 percent higher than that of SLW, and 9 percent higher than that of OPT. We believe that it is worthwhile to achieve this higher hit ratio at the cost of a small response latency increase.