Towards decentralized and adaptive network resource management

D. Tuncer*, M. Charalambides, G. Pavlou, N. Wang

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference Contribution (Conference Proceeding)

5 Citations (Scopus)

Abstract

Current practices for managing resources in fixed networks rely on off-line approaches, which can be sub-optimal in the face of changing or unpredicted traffic demand. To cope with the limitations of these off-line configurations new traffic engineering (TE) schemes that can adapt to network and traffic dynamics are required. In this paper, we propose an intra-domain dynamic TE system for IP networks. Our approach uses multi-topology routing as the underlying routing protocol to provide path diversity and supports adaptive resource management operations that dynamically adjust the volume of traffic sent across each topology. Re-configuration actions are performed in a coordinated fashion based on an in-network overlay of network entities without relying on a centralized management system. We analyze the performance of our approach using a realistic network topology, and our results show that the proposed scheme can achieve near-optimal network performance in terms of resource utilization in a responsive manner.

Original languageEnglish
Title of host publication2011 7th International Conference on Network and Service Management, CNSM 2011
Publication statusPublished - 2011
Event2011 7th International Conference on Network and Service Management, CNSM 2011 - Paris, France
Duration: 24 Oct 201128 Oct 2011

Publication series

Name2011 7th International Conference on Network and Service Management, CNSM 2011

Conference

Conference2011 7th International Conference on Network and Service Management, CNSM 2011
Country/TerritoryFrance
CityParis
Period24/10/1128/10/11

Keywords

  • Adptative Resource Management
  • Decentralized Network Configuration
  • Online Traffic Engineering

Fingerprint

Dive into the research topics of 'Towards decentralized and adaptive network resource management'. Together they form a unique fingerprint.

Cite this