A Gateway Providing Seamless Integration of TCP/IP Networks Using Satellite Links

1997-2000

Grant Reference GR/L69626/01 under the ESPRC Cds Programme

Principal Investigator

Dr G Fairhurst

Department of Engineering

University of Aberdeen

Objectives

The key goal of the project is to provide significant improvement to the throughput and delay of standard TCP/IP applications using small apperture satellite terminals. The aims of the project are summarised in the section which follows:

Abstract

A programme of research investigated design of a transparent gateway to improve the performance of the TCP transport protocol over satellite links, with particular reference to satellite links employing Satellite Digital Video Broadcast (DVB-S) links. The research will determined the elements which constrain the performance of the TCP and selected protocol algorithms which alleviated constraints on performance. The algorithms were implemented in a simulated environment to assess performance and to consider the merits and demerits of each algorithm. Interactions between existing algorithms and the new developed algorithms was investigated in the simulator, and also in practical experiments conducted using DVB-S satellite facilities provided by Globecast NE as a part of a parallel European Space Agency study. The architecture for implementation of a gateway was also considered leading to development of a set of algorithms which were implemented in an experimental system based on BSD Unix operating on a PC platform. This was used to observe the impact on the performance and reliability when used over a satellite link. A satellite link simulator was used to provide a hardware emulation of the satellite environment and to assess whether the TCP gateway may satisfy the needs of an Internet service using low cost satellite terminals. The construction and evaluation of a functional prototype gateway will be the subject of a later research programme. Since the overall concept will have been developed in this research, industrial collaboration will be sought for this subsequent development phase.

Beneficiaries

The understanding which evolves from this work may be appropriate to a range of communications users where standard protocols are unable to provide the required functionality. Two specific areas will be addressed:

  1. A commercial communications network built using Very Small Aperture Terminals. Such networks exist, but performance for many applications is limited not by the bandwidth or delay of the satellite link, but by the protocol interactions which take place. A deeper understanding of these interactions and of how to develop appropriate counter measures may have significant application - either built-in to commercial systems or implemented in a low end PC applique unit.
  2. Military users have an even more demanding need for flexible deployment. Various studies have demonstrated the real need for such equipment in a military context, but little progress has been made. The suggested approach will provide fundamental understanding, together with the architectural information needed to construct a prototype.

Additional Information

For more information contact:

Gorry Fairhurst


[List of Projects in ERG]