A version adaptive transcoding (VAT) mechanism for wireless mobile coverage networking

Abstract

Several factors affect the performance of streaming media services in the wireless mobile network including the handoff rate, high channel error rate, and available resources. In this paper, we further investigate the strategy of employing fuzzy similarities to a version adaptive transcoding (VAT) mechanism previously developed, by studying how the average startup latency, average response ratio, and average cache-hit ratio affect performance. To do this, we vary the media object size and the cache size; we also compare the approach to three classical schemes for media streaming. The goals of the VAT approach are to reduce network traffic, resulting from high network bandwidth demand and streaming constraints, and to enhance the performance of streaming media services. The proposed VAT mechanism tracks the relationship of partitioned object versions, performs the fuzzy similarity among versions and, considering the fuzzy similarity relationship, media object characteristics, cache capacity, and object version synchronization, minimizes the streaming distortion. Simulation results show that our proposed scheme has better streaming performance, in terms of such quality-of-service (QoS) metrics as average service response ratio, startup latency, and cache-hit ratio, when compared to existing server-only schemes, progressive schemes, and layered schemes.