Data broadcasting is rapidly becoming a leading candidate for the distribution of information to a large number of users in a wireless communication environment. In a broadcast data delivery system, there is a server which is broadcasting data to a user community according to a given schedule without knowing what a user needs due to the lack of communication from users to the server. A user has to wait for its needed items from the broadcast as they go by. In many applications, there are given dealines within which the user must receive the item after its need arises. If the server's broadcast schedule does not match the user's access pattern, the waiting time will be considerably long and some deadlines will not be met. If a user has a local memory, it can alleviate its deadline misses by selectively prefetching the items from the broadcast and storing them in the memory. In this paper, we identify an optimal memory management policy that minimizes the expected aggregate deadline misses. Optimal memory update strategies with limited look-ahead are given as implementable approximations of the optimal policy. Numerical experiments are performed in order to investigate the performance of the memory management strategies under different look-ahead steps, different cache sizes, different types of deadlines and different broadcast schedules.
展开▼
