TCP_STCP - Scalable TCP: Improving Performance in Highspeed...

Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon
Scalable TCP: Improving Performance in Highspeed Wide Area Networks Tom Kelly CERN - IT Division 1211 Geneva 23 Switzerland ABSTRACT TCP congestion control can perform badly in highspeed wide area networks because of its slow response with large congestion win- dows. The challenge for any alternative protocol is to better utilize networks with high bandwidth-delay products in a simple and ro- bust manner without interacting badly with existing traffic. Scal- able TCP is a simple sender-side alteration to the TCP conges- tion window update algorithm. It offers a robust mechanism to improve performance in highspeed wide area networks using tradi- tional TCP receivers. Scalable TCP is designed to be incrementally deployable and behaves identically to traditional TCP stacks when small windows are sufficient. The performance of the scheme is evaluated through experimental results gathered using a Scalable TCP implementation for the Linux operating system and a gigabit transatlantic network. The preliminary results gathered suggest that the deployment of Scalable TCP would have negligible impact on existing network traffic at the same time as improving bulk transfer performance in highspeed wide area networks. 1. INTRODUCTION A communication network can experience periods where the traf- fic offered to it exceeds the available transmission capacity; dur- ing such periods the network is said to be congested . TCP con- gestion control [9] was introduced to relieve congestion collapse that had occurred in the Internet. A result of congestion control is that resources are shared between flows during periods of con- gestion. This sharing leads to similar throughput for flows with similar round trip times and avoids starving individual flows. TCP has proved to be remarkably successful at sharing bandwidth while aggressively utilizing available capacity under a range of dynamic traffic loads. The TCP flow control algorithm uses a window and end-to-end ac- knowledgments to provide reliable data transfer across a network; Tom Kelly is a member of the Laboratory for Communication En- gineering, Cambridge University Engineering Department, Trump- ington Street, Cambridge CB2 1PZ, United Kingdom. a brief description is given here and a more complete reference is [15]. The sending host maintains a congestion window, cwnd , which places an upper bound on the number of segments that may be sent into the network awaiting acknowledgment by the receiver. 1 Upon receiving a data packet the receiver schedules a cumulative acknowledgment, that covers all received packets, to be sent to the sender. The receiver also advertises to the sender a receive window, rwnd , which is the size of the available socket receive buffer for this connection. The sender is allowed to have at most the minimum of cwnd and rwnd packets in the network awaiting acknowledg- ment. The receive window provides flow control for the receiving application; if the receiving application cannot process data at the
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Image of page 2
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 01/07/2011 for the course CSE 100 taught by Professor Staff during the Spring '08 term at UCSD.

Page1 / 8

TCP_STCP - Scalable TCP: Improving Performance in Highspeed...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online