Super-Streaming: A New Object Delivery Paradigm forContinuous Media Servers

  • Authors:

  • Cyrus Shahabi;Mohammad H. Alshayeji

  • Affiliations:

  • Integrated Media Systems Center and Computer Science Department, University of Southern California, Los Angeles, California 90089, USA. cshahabi@cs.usc.edu;Integrated Media Systems Center and Computer Science Department, University of Southern California, Los Angeles, California 90089, USA. alshayej@cs.usc.edu

  • Venue:
  • Multimedia Tools and Applications
  • Year:
  • 2000

Quantified Score

Hi-index 0.00

Visualization

Abstract

A number of studies have focused on the design ofcontinuous media, CM, (e.g., video and audio) servers to support thereal-time delivery of CM objects. These systems have been deployedin local environments such as hotels, hospitals and cruise ships tosupport media-on-demand applications. They typically streamCM objects to the clients with the objective of minimizing the bufferspace required at the client site. This objective can now be relaxeddue to the availability of inexpensive storage devices at the clientside. Therefore, we propose a Super-streaming paradigm thatcan utilize the client side resources in order to improve theutilization of the CM server. To support super-streaming, we proposea technique to enable the CM servers to deliver CM objects at a ratehigher than their display bandwidth requirement. We also proposealternative admission control policies to downgrade super-streams infavor of regular streams when the resources are scarce. Wedemonstrate the superiority of our paradigm over streaming with bothanalytical and simulation models.Moreover, new distributed applications such as distant-learning,digital libraries, and home entertainment require the delivery of CMobjects to geographically disbursed clients. For quality purposes,recently many studies proposed dedicated distributed architectures tosupport these types of applications. We extend our super-streamingparadigm to be applicable in such distributed architectures. Wepropose a sophisticated resource management policy to supportsuper-streaming in the presence of multiple servers, network linksand clients. Due to the complexity involved in modeling thesearchitectures, we only evaluate the performance of super-streaming bya simulation study.