By Topic

A new self-organizing approach for congestion control in ATM networks

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$33 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

2 Author(s)
Mohamed, A.A. ; Dept. of Electr. & Comput. Eng., Dalhousie Univ., Halifax, NS, Canada ; El-Hawary, M.

This paper presents a new approach to the problem of congestion control arising at a non-blocking ATM switch with input queues. When two or more packets in the switch input buffers simultaneously request the same output port of an ATM switch, congestion will occur, and one of these packets will be transmitted and the others will be either buffered or dropped. In order to prevent this cell discarding, input queuing is widely used. The only problem with input queuing is the head-of-line (HOL) blocking in which a packet in a first-in-first-out (FIFO) queue has no chance to access an available output port because the packet ahead of it in the buffer is blocked. In order to solve this problem, many algorithms have been presented. One of these algorithms is the input window policy algorithm which divides the contention resolution process into S steps during a time slot to find an optimal set of packets. In the optimal set, each packet must have a different destination address. We introduce a new self-organizing neural network approach to resolve the contention problem based on the input window policy algorithm. From the simulation results shown, our neural network model provides the same switch throughput as that of the input window policy. However, the new approach is definitely much faster than the original algorithm in terms of finding this optimal set due to the parallel processing performance of the neural networks

Published in:

Electrical and Computer Engineering, 2000 Canadian Conference on  (Volume:2 )

Date of Conference: