E-Speed start - A window based additive increase, multiplicative decrease network congestion control technique

0Citations
Citations of this article
2Readers
Mendeley users who have this article in their library.
Get full text

Abstract

The two major algorithms in the Additive Increase, Multiplicative Decrease Network Congestion Control Algorithm (AIMD) are the slow start and the congestion avoidance. Much works has been done on improving the congestion avoidance stage. Not until recently has attention shifted to slow start. The proposals for improving slow start include swift start, quick start etc. These modifications are deficient in one way or the other. Hence we propose e-speed start for AIMD. We attempt to incorporate environmental factors into slow start speed determination. Furthermore, we propose multiple selectable startups for AIMD. Hence a connection may use the conventional slow start, e-speed start or any other startup algorithm depending on the size of the available bandwidth. In fact, we are of the view that AIMD slow start may be selectable from n-ary set of algorithms. The e-speed start uses the available bandwidth to calculate a factor β which is used to determine the congestion window (cwnd) suitable for transmission at that point. The result obtained shows that e-speed start performs far better than the slow start where the available bandwidth is high and the environmental factors are favourable. It was observed that e-speed start is able to increase throughput between 700-51100% and network link utilisation by 37.5% as against the convectional slow-start technique in TCP. © Springer Science+Business Media B.V. 2010.

Cite

CITATION STYLE

APA

Oyeyinka, I. K., Akinwale, A. T., Oluwatope, A. O., Folorunso, O., Aderounmu, G. A., & Abiona, O. O. (2010). E-Speed start - A window based additive increase, multiplicative decrease network congestion control technique. In Technological Developments in Networking, Education and Automation (pp. 493–498). Kluwer Academic Publishers. https://doi.org/10.1007/978-90-481-9151-2_85

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Save time finding and organizing research with Mendeley

Sign up for free