Rate-Based Congestion Control in Networks with Smart Links

Abstract

The author uses a network simulator to explore rate-based congestion control in networks with smart links that can feed back information to tell senders to adjust their transmission rates. This method differs in a very important way from congestion control in which a congested network component just drops packets - the most commonly used method. It is clearly advantageous for the links in the network to communicate with the end users about the network capacity, rather than the users unilaterally picking a transmission rate. The components in the middle of the network, not the end users, have information about the capacity and traffic in the network. The author experiments with three different algorithms for calculating the control rate to feed back to the users. All of the algorithms exhibit problems in the form of large queues when simulated with a configuration modeling the dynamics of a packet-voice system. However, the problems are not with the algorithms themselves, but with the fact that feedback takes time. If the network steady-state utilization is low enough that it can absorb transients in the traffic through it, then the large queues disappear. If the users are modified to start sending slowly, to allow the network to adapt to a new flow without causing congestion, a greater portion of the network's bandwidth can be used.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jan 01, 1990
Accession Number
ADA219852

Entities

People

  • Andrew T. Heybey

Organizations

  • Massachusetts Institute of Technology

Tags

Communities of Interest

  • Space

DTIC Thesaurus Topics

  • Classification
  • Computer Communications
  • Computer Networks
  • Computer Science
  • Contracts
  • Control Systems
  • Data Transmission
  • Electrical Engineering
  • Engineering
  • Hypervelocity Flow
  • Information Processing
  • Military Research
  • Network Science
  • Security
  • Simulations
  • Simulators
  • Transport Protocols

Fields of Study

  • Computer science

Readers

  • Mathematical Modeling and Probability Theory.
  • Radio communications and signal processing.
  • Strategic Security Studies