Performance optimization problem in speculative prefetching

Abstract

Speculative prefetching has been proposed to improve the response time of network access. Previous studies in speculative prefetching focus on building and evaluating access models for the purpose of access prediction. This paper investigates a complementary area which has been largely ignored, that of performance modeling. We analyze the performance of a prefetcher that has uncertain knowledge about future accesses. Our performance metric is the improvement in access time, for which we derive a formula in terms of resource parameters (time available and time required for prefetching) and speculative parameters (probabilities for next access). We develop a prefetch algorithm to maximize the improvement in access time. The algorithm is based on finding the best solution to a stretch knapsack problem, using theoretically proven apparatus to reduce the search space. An integration between speculative prefetching and caching is also investigated.

Multimedia prefetching with optimal Markovian policies

“…The approach developed in Tuah et al (2002) provides a sequential policy aiming to optimize the latency at each navigation step. In their work, the authors assume the knowledge of the available bandwidth, the size of the media objects, and the time spent by the user in each object.…”

“…A second approach aiming to reach the optimality for the sequential policy proposed in Tuah et al (2002) is presented by Angermann (2002). The prefetching mechanism uses a Markov model that is downloaded to the client's system that becomes personalized to the client's distinct behavior.…”

“…The prefetching mechanism uses a Markov model that is downloaded to the client's system that becomes personalized to the client's distinct behavior. The framework used is slightly different from Tuah et al (2002) and allows the media objects to be stored on different servers, the cache storage is supposed unlimited and the viewing time for each object is not anymore constant but distributed following a Zipf low (Breslau et al, 1999). Under these hypothesis, the author shows that the media objects should be prefetched in decreasing order of their access probabilities.…”

Optimal Prefetching Scheme in P2P VoD Applications With Guided Seeks

“…Prefetching is an effective approach to reduce the waiting time for an object in the applications of file transfers and media streaming. In [19], the authors optimize the speculative prefetching scheme by solving a stretched knapsack problem. In [20], a prefetching algorithm with low computational complexity is proposed to minimize the expected waiting time of document retrieval.…”