On performance objective functions for optimising routed networks for best QoS

Pham, Huan, and Lavery, Bill (2003) On performance objective functions for optimising routed networks for best QoS. In: Proceedings of Quality of Service in Multiservice IP Networks: Second International Workshop, QoS-IP 2003 (2601), pp. 64-75. From: Quality of Service in Multiservice IP Networks: Second International Workshop, QoS-IP 2003, 24-26 February 2003, Milano, Italy.

[img] PDF (Published Version) - Published Version
Restricted to Repository staff only

View at Publisher Website: http://www.springerlink.com/content/9cbq...
 
5


Abstract

This paper contributes to understanding of how Quality of Service (QoS) can be optimised in Autonomous System (AS) networks using Open Shortest Path First (OSPF) routing. We assume that, although guaranteed QoS can not be provided, network wide performance can be optimised in a manner which significantly improves QoS. We particularly focus on devising a network performance objective, which is convenient for the network optimization process, while closely representing QoS experience for end users. Our results show that, firstly, network-wide performance can be significantly improved by optimizing OSPF weights. Secondly, the resulting optimal set of weights and network performance depends very much on the definition of the optimization objective function. We argue that min max utilisation is an appropriate function for heavily loaded network, while end to end delay is more appropriate optimization objective for lightly loaded condition, at which most of the backbones are operating.

Item ID: 7821
Item Type: Conference Item (Refereed Research Paper - E1)
Keywords: autonomous system networks; quality of service
Related URLs:
ISSN: 1611-3349
Date Deposited: 22 Jul 2010 04:16
FoR Codes: 09 ENGINEERING > 0906 Electrical and Electronic Engineering > 090609 Signal Processing @ 100%
SEO Codes: 89 INFORMATION AND COMMUNICATION SERVICES > 8901 Communication Networks and Services > 890199 Communication Networks and Services not elsewhere classified @ 100%
Downloads: Total: 5
More Statistics

Actions (Repository Staff Only)

Item Control Page Item Control Page