Optimising Quality of Service through the controlled aggregation of traffic

ME Culverhouse, BV Ghita, P Reynolds, X Wang

Research output: Contribution to journalConference articlepeer-review

Abstract

The multimedia enriched transport supported by the Internet today continues to pose a challenge to engineers who attempt to provide QoS for its users. The popular method of using Diffserv solutions is typically too static to meet the mixed user traffic model. This paper introduces a novel user-centric approach of dynamically evaluating and policing incoming Internet flows to control the ratio of traffic types for individual users. After describing its theoretical rationale, the proposed method is presented as an integrated architecture (Congestion Aware Packet Scheduler - CAPS), which allows seamless integration with existing Diffserv networks. An ns2 implementation of the CAPS architecture is presented and investigated for a number of different scenarios. The evaluation of CAPS indicates that the architecture outperforms best effort, traditional Diffserv and weighted-RED alternatives, providing a better, dynamic QoS balance for a wide range of traffic profiles without the need for explicit predetermined precedence on traffic types.
Original languageEnglish
Pages (from-to)1-7
Number of pages0
JournalInternet Technology and Secured Transactions (ICITST), 2010 International Conference for DOI -
Volume0
Issue number0
Publication statusPublished - 2010
EventInternet Technology and Secured Transactions (ICITST), 2010 International Conference for -
Duration: 1 Jan 2010 → …

Keywords

  • Internet
  • computer networks
  • multimedia communication
  • optimisation
  • quality of service
  • scheduling
  • telecommunication congestion control
  • telecommunication traffic
  • Diffserv solutions
  • congestion aware packet scheduler
  • controlled traffic aggregation
  • multimedia enriched transport
  • quality of service optimisation

Fingerprint

Dive into the research topics of 'Optimising Quality of Service through the controlled aggregation of traffic'. Together they form a unique fingerprint.

Cite this