Abstract
One reason for the success of TCP and its widespread usage in the Internet is its ability to control network congestion. A TCP source uses implicit notification of network congestion (via packet loss) to control the rate at which it sends data. The rate is controlled by increasing or decreasing the window of outstanding, unacknowledged data. In particular, TCP uses “additive increase and multiplicative decrease” to increase the rate linearly during times of no packet loss, and to decrease the rate geometrically when loss occurs. As a consequence, the rate at which a TCP source transmits tends to be periodic over time. For example, Figure 1 shows the periodic behavior of a single TCP source in an otherwise idle network.
This work was done in collaboration with Youngmi Joo (1970-2001). It is dedicated to her memory.
This is a preview of subscription content, log in via an institution.
Preview
Unable to display preview. Download preview PDF.
Author information
Authors and Affiliations
Editor information
Rights and permissions
About this chapter
Cite this chapter
Gilbert, A.C. Dynamics of Congestion Control. In: Kocarev, L., Vattay, G. (eds) Complex Dynamics in Communication Networks. Understanding Complex Systems. Springer, Berlin, Heidelberg. https://doi.org/10.1007/10973509_2
Download citation
DOI: https://doi.org/10.1007/10973509_2
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-24305-2
Online ISBN: 978-3-540-32381-5
eBook Packages: Physics and AstronomyPhysics and Astronomy (R0)