Browse Prior Art Database

Fault Tolerance through Replication of Video Assets

IP.com Disclosure Number: IPCOM000118077D
Original Publication Date: 1996-Sep-01
Included in the Prior Art Database: 2005-Mar-31
Document File: 4 page(s) / 155K

Publishing Venue

IBM

Related People

Harter, JL: AUTHOR [+3]

Abstract

Disclosed is a method of improving the reliability of delivering video data streams (from a file server to other machines or devices in a communications network) through the use of duplicate copies of video assets. The method provides a degree of fault tolerance from various types of device failures by diverting requests for video data on those devices to alternate devices that contain the same video data. While this same effect can be achieved with other hardware (mirroring) and software (redundant file systems), this method can provide some advantages over those implementations.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 37% of the total text.

Fault Tolerance through Replication of Video Assets

      Disclosed is a method of improving the reliability of
delivering video data streams (from a file server to other machines
or devices in a communications network) through the use of duplicate
copies of video assets.  The method provides a degree of fault
tolerance from various types of device failures by diverting requests
for video data on those devices to alternate devices that contain the
same video data.  While this same effect can be achieved with other
hardware (mirroring) and software (redundant file systems), this
method can provide some advantages over those implementations.

This method depends on the existence of a file server with the
following characteristics:

      The file server has access to a large amount of Direct Access
Storage Devices (DASD) that contain digital representations of video
and audio data, such as movies, television programs, news broadcasts,
music videos, etc..  Associated with each of these video/audio assets
is certain quality-of-service information that describes how fast the
asset's data must be read from these DASD devices and delivered to a
video/audio device in order to reproduce a quality picture/sound.
Failure to deliver data at this rate results in disruption of the
video/audio signal...  in other words, poor picture/sound quality.
For this reason, the file server should be capable of "pre-reading"
video/audio data from DASD into memory, in order to minimize the
impact of DASD latency on the delivery of the video/audio data.  By
performing this "aggressive read-ahead" of video/audio data (so that
the data is already in memory when the request for the data arrives),
the file server can continue delivering data at the required rate,
even if DASD errors are causing the file server to perform retries.

      Note: For simplification, the remainder of this article will
refer to the data on these DASD as video data and video assets, even
though the data may be of any type.

      A communication subsystem that allows the file server to send
this digital video data to other devices in the network, call them
video end-points.  The communication subsystem is assumed to be fast
enough to accommodate the quality-of-service (data rate) requirements
of the video assets.

      A mechanism of recording the capacities (size and speed) of
these DASD and communication devices, and monitoring and controlling
the load against them in such a way as to guarantee that the
capacities will not be exceeded.  This mechanism allows the server to
continue accepting requests (from video end-points) for video data
streams as long as the DASD and communication devices can accommodate
the load, and to deny any request for a video data stream that would
exceed the capacity of any one of the involved devices.

      In order to accommodate more requests for a single video asset
than the DASD (on which the video asset resides) can support, the
f...