A novel server-side proxy caching strategy for large-scale multimedia applications

Nowadays, server-side Web caching becomes an important technique used to reduce the User Perceived Latency (UPL). In large-scale multimedia systems, there are many Web proxies, connected with a multimedia server, that can cache some most popular multimedia objects and respond to the requests for them. Multimedia objects have some particular characteristic, e.g., strict QoS requirements. Hence, even some efficient conventional caching strategies based on cache hit ratio, meant for non-multimedia objects, will confront some problems in dealing with the multimedia objects. If we consider additional resources of proxy besides cache space, say bandwidth, we can readily observe that high hit ratios may deteriorate the entire system performance. In this paper, we propose a novel placement model for networked multimedia systems, referred to as the Hk/THk/T model, which considers the combined influence of arrival rate, size, and playback time to select the objects to be cached. Based on this model, we propose an innovative Web caching algorithm, named as the ART-Greedy algorithm, which can balance the load among the proxies and achieve a minimum Average Response Time (ART) of the requests. Our experimental results conclusively demonstrate that the ART-Greedy algorithm outperforms the most popular and commonly used LFU (Least Frequently Used) algorithm significantly, and can achieve a better performance than the byte-hit algorithm when the system utilization is medium and high.

Research highlights
► Server-side Web caching is an important technique used to reduce the User Perceived Latency.
► Multimedia objects have some particular characteristics, and conventional cache algorithms work poorly.
► We propose and use Hk/THk/T model, which considers the combined influence of arrival rate, size, and playback time to select the objects to be cached.
► We propose ART-Greedy algorithm and show that it outperforms the most popular and commonly used LFU (Least Frequently Used) algorithm significantly.