Browse Prior Art Database

Predicting Congestion in Packet Switched Networks

IP.com Disclosure Number: IPCOM000084430D
Original Publication Date: 1975-Nov-01
Included in the Prior Art Database: 2005-Mar-02
Document File: 2 page(s) / 29K

Publishing Venue

IBM

Related People

Friedman, SW: AUTHOR [+3]

Abstract

IBM Technical Disclosure Bulletin, Vol. 18, No. 6, November 1975, page 1784, describes path selection in packet switched data communication networks, based upon interchange of accurate and timely path delay information by nodal exchange stations.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 77% of the total text.

Page 1 of 2

Predicting Congestion in Packet Switched Networks

IBM Technical Disclosure Bulletin, Vol. 18, No. 6, November 1975, page 1784, describes path selection in packet switched data communication networks, based upon interchange of accurate and timely path delay information by nodal exchange stations.

A problem inherent in utilization of such networks is that by the time actual congestion is detected and communicated, the affected facilities may be overloaded and the affected traffic "log-jammed" to the point that faulty acknowledgments and associated repeat transmissions are generated which tend to intensify the congestion.

A technique presently suggested, which avoids or reduces this problem, is to employ predictive delay measurement and communication procedures throughout the network, by which the network loads can be anticipatorily balanced before congestion reaches the critical log-jamming level just described. The accompanying graph illustrates a basis for suitable predictive delay measurement at any node.

The following expression is represented: Lq = R/2/ Over 1-r where, Lq = Expected traffic (queue length). r = Actual occupancy (in %) of nodal facilities.

Clearly, in the critical region small percent changes in actual occupancy are accompanied by comparatively large changes in anticipated traffic. Thus, by frequently measuring changes in traffic inflow and queue occupancy, the critical region condition can be anticipated and communicated throughout the network ...