Research: Techniques for multimedia synchronization in network file systems

  • Authors:
  • P Venkat Rangan;Srinivas Ramanathan;Harrick M Vin;Thomas Kaeppner

  • Affiliations:
  • Multimedia Laboratory, Department of Computer Science and Engineering, University of California at San Diego, La Jolla, CA 92093-0114, USA;Multimedia Laboratory, Department of Computer Science and Engineering, University of California at San Diego, La Jolla, CA 92093-0114, USA;Multimedia Laboratory, Department of Computer Science and Engineering, University of California at San Diego, La Jolla, CA 92093-0114, USA;Multimedia Laboratory, Department of Computer Science and Engineering, University of California at San Diego, La Jolla, CA 92093-0114, USA

  • Venue:
  • Computer Communications
  • Year:
  • 1993

Quantified Score

Hi-index 0.24

Visualization

Abstract

One of the unique features that distinguishes digital multimedia from traditional computer data is the presence of multiple media streams, whose display must proceed in a mutually synchronized manner. The design of techniques for synchronization of multimedia data at the time of storage, and retrieval from network file servers is the subject matter of this paper. We present algorithms by which a file server can create a relative time system and synchronize media units transmitted by different sources on a network to construct a multimedia object. These algorithms stay robust in the absence of global clocks, presence of transmission jitter and generation rate mismatches. We develop a feedback technique using which the file server can detect asynchronies in display devices during retrieval of multimedia objects, and even restore synchrony by deleting or duplicating media units destined for asynchronous destinations. We then present strategies by which the file server can actually predict the time in future when the asynchrony of a device is expected to exceed the permitted bound, and take gradual preventive action to nullify the asynchrony in advance. These algorithms can be generalized to heterogeneous multimedia networks in which there may be variations in sizes of media units generated, differences in network locations of sources and destinations, etc. We are currently implementing these techniques in a digital multimedia on-demand storage server being developed at the UCSD Multimedia Laboratory.