On the Extensibility Properties and Performance Measures of Circuit Switched Telecommunication Networks, Using Agent-Based Distributed Routing Algorithm
Date Issued
2003
Abstract
An agent-based routing control system and its behavior
while network topology changes to an extended version, is
being studied in this paper. In real time networks for various
reasons some of the nodes collapse or taken off-line for
maintenance causing a significant reduction in performance.
Inspired from this, we have examined the behavior of an
agent-based routing which is applied to a realistic
Synchronous Digital Hierarchy (SDH) complex topology
network that tends to increase-or decrease- the number of
nodes randomly/under conditions or even activate certain
nodes that were inactive. Algorithmically complex problems
like routing in telecommunication network, needs a
dynamic adaptive approach. Agent-based routing method
continually modifies the routing tables in response to
congestion and overload in a network, to two different
network topologies, measuring performance for the
assumptions made. This decentralized routing scheme is
shown not be substantially affected from any topological
changes, when subjected to a certain frequency limit in
variations. Through various newly introduced metrics and
simulation of the proposed scenarios, we have shown that
agent-based routing offer us a decentralized routing control
in the network and an efficient way to stabilize
responsiveness and increase overall performance.
while network topology changes to an extended version, is
being studied in this paper. In real time networks for various
reasons some of the nodes collapse or taken off-line for
maintenance causing a significant reduction in performance.
Inspired from this, we have examined the behavior of an
agent-based routing which is applied to a realistic
Synchronous Digital Hierarchy (SDH) complex topology
network that tends to increase-or decrease- the number of
nodes randomly/under conditions or even activate certain
nodes that were inactive. Algorithmically complex problems
like routing in telecommunication network, needs a
dynamic adaptive approach. Agent-based routing method
continually modifies the routing tables in response to
congestion and overload in a network, to two different
network topologies, measuring performance for the
assumptions made. This decentralized routing scheme is
shown not be substantially affected from any topological
changes, when subjected to a certain frequency limit in
variations. Through various newly introduced metrics and
simulation of the proposed scenarios, we have shown that
agent-based routing offer us a decentralized routing control
in the network and an efficient way to stabilize
responsiveness and increase overall performance.