Browse Prior Art Database

Method to Deliver Scalable Video across a Distributed Computer System

IP.com Disclosure Number: IPCOM000112441D
Original Publication Date: 1994-May-01
Included in the Prior Art Database: 2005-Mar-27
Document File: 6 page(s) / 302K

Publishing Venue

IBM

Related People

Baugher, M: AUTHOR [+2]

Abstract

A method is disclosed for delivering multimedia data across a network, even in the presence of congestion. The method involves scaling back of I/O requests based on a 'colored index' access method when network congestion is detected.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 18% of the total text.

Method to Deliver Scalable Video across a Distributed Computer System

      A method is disclosed for delivering multimedia data across a
network, even in the presence of congestion.  The method involves
scaling back of I/O requests based on a 'colored index' access method
when network congestion is detected.

      Transient congestion of distributed system components may
disrupt multimedia applications:  Sound, voice, and video flows
require throughput, delay, and delay variation guarantees.
Multimedia applications need an adaptation mechanism to cope with
short term congestion in a distributed computer system.  The
mechanism must "scale down" a multimedia flow by dropping
"nonessential" transmissions, e.g., delta video frames to better
ensure the delivery of "essential" transmissions, e.g., reference
video frames and audio.  Scaling back the flow also temporarily
reduces distributed system load until congestion subsides and the
multimedia flow may be "scaled up" again.  The needed solution must
provide scalable video transfer across existing file systems.  The
solution should work with today's "client pull" file systems as well
as for future "server push" implementations.

      Multimedia applications may use remote devices such as disks,
cameras, speakers and video windows.  When multimedia flows have
remote sources or sinks, they require some Quality of Service (QoS)
guarantees across the distributed system, such as reserved bandwidth
across a computer network.  Today's packet networks can at best
provide statistical guarantees since bursty data traffic may
unpredictably delay multimedia flows.  When delivery deadlines are
missed, multimedia elements are effectively lost.

      Some multimedia streams tolerate loss, especially when it is
controlled.  It's a happy coincidence that the medium which is most
tolerant to loss is video, the medium requiring the greatest
bandwidth from a network.  The Motion Picture Experts' Group's
(MPEG's) current video decompression and streaming standard, ISO
11172, identifies three "scales" that can be used for reducing the
size of a video stream:  Removal of spatial redundancy permits an
encoding of repeated elements many of which can be dropped at the
source and reconstructed at the destination (removal of "white
space"), and in some cases information can be imperceptibly dropped
by manipulating the brightness and color of the frame; removal of
temporal redundancy permits the dropping of unchanging elements from
the successive frames, and motion compensation permits the dropping
of elements in change during motion by performing interpolation at
the destination.  ISO 11172 defines I pictures as frames that are
complete save for the removal of spatial redundancy, P pictures as
frames which depend upon I pictures or other P pictures to permit the
removal of temporal redundancy, and B pictures as frames sandwiched
between I or P pictures that have motion compensatable redundan...