Fine-grained scalable video caching for heterogeneous clients

J. Liu*, J. Xu, X. Chu

*Corresponding author for this work

Research output: Contribution to journalJournal articlepeer-review

9 Citations (Scopus)
31 Downloads (Pure)

Abstract

Much research has focused on caching adaptive videos to improve system performance for heterogeneous clients with diverse access bandwidths. However, existing rate-adaptive caching systems, which are based on layered coding or transcoding, often suffer from a coarse adaptation and/or a high computation overhead. In this paper, we propose an innovative rate-adaptive caching framework that enables low-cost and fine-grained adaptation by using MPEG-4 fine-grained scalable videos. The proposed framework is both network-aware and media-adaptive; i.e., the clients can be of heterogeneous streaming rates, and the backbone bandwidth consumption can be adaptively controlled. We develop efficient cache management schemes to determine the best contents to cache and the optimal streaming rate to each client under the framework. We demonstrate via simulations that, compared to nonadaptive caching, the proposed framework with the optimal cache management not only achieves a significant reduction in the data transmission cost, but also enables a flexible utility assignment for the heterogeneous clients. Our results also show that the framework maintains a low computational over-head, which implies that it is practically deployable.

Original languageEnglish
Article number1703515
Pages (from-to)1011-1019
Number of pages9
JournalIEEE Transactions on Multimedia
Volume8
Issue number5
DOIs
Publication statusPublished - Oct 2006

Scopus Subject Areas

  • Signal Processing
  • Media Technology
  • Computer Science Applications
  • Electrical and Electronic Engineering

User-Defined Keywords

  • Fine-grained scalable video
  • Proxy caching
  • Resource allocation
  • Streaming media

Fingerprint

Dive into the research topics of 'Fine-grained scalable video caching for heterogeneous clients'. Together they form a unique fingerprint.

Cite this