Scheduling of storage and cache servers for replicated multimedia data

Multimedia storage servers will play an important role in many forthcoming network-based multimedia services like video-on-demand (VoD). The load on the storage servers and the data transmission cost can be reduced by using additional cache servers which store the data requested by a user temporarily and provide data to other users located near the cache servers. It is required to minimize the sum of the storage cost and the transmission cost while satisfying users' timing requirements. We study a scheduling scheme for storage servers and cache servers to meet this requirement. Assuming that the locations of the data and the users' deadline requirements are known beforehand, we formulate the problem, devise a greedy heuristic algorithm for scheduling servers statically and analyze its complexity, and, by limiting the backtracking distance and restricting the search area using a constant-size search window, the scheduling time can be greatly reduced with little increase in the total cost.