The Utlization of Distributed Processing on Signaling Calls in High Speed ATM Network: Architecture, Modeling, Simulation, Performance Analysis

摘要

This paper presents an efficient distributed call processingrnarchitecture to accelerate the call setup process in ATM networks.rnAlthough distributed call processing had been previouslyrnemployed in classical telephony, the study reportedrnhere is unique in that, in ATM networks, call processing isrndriven by user call requests and Quality of Service (QoS)rnmetrics. The architecture is designed to address a highlyrnstressed network, just short of driving it into “instability”.rnTwo representative networks – a 15-node and a 50-nodernATM network spanning the continental US, are modeledrnin an asynchronous distributed simulator and executed onrnan asynchronous distributed testbed of 45+ linux workstations.rnThe underlying distributed hardware architecturesrnfor the ATM nodes are modeled utilizing realistic values forrnthe logic and timing parameters. Extensive simulation experimentsrnare designed and executed to yield insights intornnetwork performance behavior, encapsulated through keyrnmetrics including call success rate, call setup time, and linkrnbandwidth utilization, as a function of the important parametersrnof the distributed call processing architecture includingrnthe number of call processors at a node, the degree ofrncontention for the shared memory, and shared memory accessrnspeed.