Browse Prior Art Database

Window Based Bandwidth Adjustment

IP.com Disclosure Number: IPCOM000113761D
Original Publication Date: 1994-Oct-01
Included in the Prior Art Database: 2005-Mar-27
Document File: 2 page(s) / 72K

Publishing Venue

IBM

Related People

Galand, G: AUTHOR [+4]

Abstract

The traffic generated by a terminal equipment is sent to the port of a high speed network for transmission over a wide area network. At the port, the traffic is regulated by a leaky bucket circuit which is characterized by a set of parameters: the token generation rate and the token pool size. The token generation rate is tuned to the bandwidth that has been reserved in the network for this traffic, and the token pool size is tuned to the maximum burst size. Traffic in excess is tagged for further discard. This mechanism ensures that the network is protected against any traffic that would be generated in excess by the terminal equipment.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 53% of the total text.

Window Based Bandwidth Adjustment

      The traffic generated by a terminal equipment is sent to the
port of a high speed network for transmission over a wide area
network.  At the port, the traffic is regulated by a leaky bucket
circuit which is characterized by a set of parameters:  the token
generation rate and the token pool size.  The token generation rate
is tuned to the bandwidth that has been reserved in the network for
this traffic, and the token pool size is tuned to the maximum burst
size.  Traffic in excess is tagged for further discard.  This
mechanism ensures that the network is protected against any traffic
that would be generated in excess by the terminal equipment.

      It may be desirable to adapt the parameters of the leaky bucket
to an evolving traffic.  For this purpose, a generic approach to
bandwidth estimation and adaptation has been proposed in [*].  This
approach is based on traffic monitoring and exponential modeling, and
allows to determine the leaky bucket parameters from the knowledge of
the mean bit rate and the tagging probability.  Described is a window
based implementation of the method.

      At the leaky bucket admission buffer, the mean input rate m,
and the tagging probability xi are measured at each sampling time t=1
second.  The filtered values  m hat sub t  and  xi hat sub t
 are estimated and observed during a 30 second time window.  At the
end of this window the following values are computed:
  m hat sub mean = 1 over < Card midsub < t memberof window >
  lbrace m hat sub t rbrace > %% sum from < t memberof window >
     ...