Routing With Topology Aggregation in Delay-Bandwidth Sensitive Networks

King Shan Lui, Klara Nahrstedt, Shigang Chen

Research output: Contribution to journalArticlepeer-review


Routing is a process of finding a network path from a source node to a destination node. The execution time and the memory requirement of a routing algorithm increase with the size of the network. In order to deal with the scalability problem, large networks are often structured hierarchically by grouping nodes into different domains. The internal topology of each domain is then aggregated into a simple topology that reflects the cost of routing across that domain. This process is called topology aggregation. For delay-bandwidth sensitive networks, traditional approaches represent the property of each link in the aggregated topology as a delay-bandwidth pair, which corresponds to a point on the delay-bandwidth plane. Since each link after aggregation may be the abstraction of many physical paths, a single delay-bandwidth pair results in significant information loss. The major contribution of this paper is a novel quality-of-service (QoS) parameter representation with a new aggregation algorithm and a QoS-aware routing protocol. Our QoS representation captures the state information about the network with much greater accuracy than the existing algorithms. Our simulation results show that the new approach achieves very good performance in terms of delay deviation, success ratio, and crankback ratio.

Original languageEnglish (US)
Pages (from-to)17-29
Number of pages13
JournalIEEE/ACM Transactions on Networking
Issue number1
StatePublished - Feb 2004


  • Delay-bandwidth sensitive networks
  • Hierarchical routing
  • QoS routing
  • Topology aggregation

ASJC Scopus subject areas

  • Software
  • Computer Science Applications
  • Computer Networks and Communications
  • Electrical and Electronic Engineering

Fingerprint Dive into the research topics of 'Routing With Topology Aggregation in Delay-Bandwidth Sensitive Networks'. Together they form a unique fingerprint.

Cite this