Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.649050
Title: Integer-time burst-level simulation techniques for very high speed communication networks
Author: Cusack, Stephen D.
Awarding Body: University of Edinburgh
Current Institution: University of Edinburgh
Date of Award: 2000
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
Abstract:
This thesis presents a technique which abstracts the level of detail of a cell-based network to the burst level, where a burst describes a group of cells in transmission. the use of the burst is designed to reduce the number of events which must be processed to perform a network simulation. To leverage the power of today's inexpensive, high performance microprocessors, the techniques presented exclusively use all-integer arithmetic. The use of integer arithmetic provides an inherent performance gain over floating point arithmetic. The use of 64-bit integer arithmetic is very desirable, which is a completely realistic goal for the next generation of microprocessors. Techniques using integer arithmetic to multiplex, queue, demultiplex and switch bursts of cells are presented. Each operation is presented as a simulation object integrated into an efficient C++-based object-oriented bespoke simulation environment. The accuracy and performance issues for each object are explored, in comparison with an efficient cell level simulator also developed in the work. Detailed investigation of the proposed techniques highlights two core operations, which are then further optimised as integer techniques (by removing the integer divide operation). The revised integer techniques are shown to improve the performance of the simulation objects, while preserving the accuracy of the techniques. Three basic experiments are presented in the thesis to show how non-trivial simulations can be constructed from the core simulation objects presented. The performance and accuracy implications of each experiment are analysed and used to provide guidance for further work based on the techniques presented.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.649050  DOI: Not available
Share: