Cache Replacement Scheme Based on Sliding Window and TTL for Video on Demand

In this paper, we propose a novel two-phased cache replacement algorithm based on the dataset obtained from a typical wireless mesh network. In the first phase, we aim to make a trade-off between the networks resources and refresh the cache content. To deal with this issue, we apply a sliding window to propose a new parameter, called equivalent active time, to forecast the user's behaviour pattern, i.e., make a decision whether to update the cache contents or utilize the network's resources for other services. In the second phase, we introduce, for the first time, the parameter time to live (TTL) that shows the video popularity lifetime. Our proposed cache replacement strategy uses frequency counters and the TTL information in the victim selection process to prevent the cache pollution and make better use of the cache space. Numerical results show that our replacement algorithm outperforms some existing cache replacement strategies in term of a portion of satisfied requests.