Rank-based Contents Caching and Distribution Scheme for Resource-Efficient Fast Channel Zapping in IPTV Service

In this paper, we propose a rank-based contents caching and distribution scheme of caching channels in FCZ (Fast Channel Zapping) cache servers. Cache servers store certain amount of recent multicast streams including intra frame for all channels and serve cached data at high speed in unicast stream while a set-top box requests cached data regarding the newly joined multicast stream on channel change. The set-top box can achieve the fast channel change by reassembling the cached data only with a small amount of multicast stream reception for new channel. However, most of channels under caching in a cache server have not been chosen because channel change toward non-popular channels rarely happens. We address here how caching channels are efficiently distributed over cache servers in server's resource point of view. Since viewer rating of IPTV channels is biased to a few popular channels, cache servers can take channel popularity into consideration for optimizing overall resource utilization. The proposed scheme classifies IPTV channels into two groups, popular channel group and non-popular channel group, based on popularity rank. Each cache server is then assigned to the appropriate channel group, which not only saves its resources by caching the data from its assigned channel, but also improves serviceability (the number of concurrent FCZ sessions), i.e., not necessary caching all the channels alone. Test results prove superiority of the proposed scheme in terms of the total number of concurrent FCZ sessions.