next up previous
Next: Congestion Control Up: Improving TCP Congestion Control Previous: Previous Work

   
TCP Santa Cruz - Protocol Description

Our goal in designing this protocol was to work within the existing TCP specifications, but to alter the protocol in such a way as to utilize network bandwidth without the type of congestion-causing probing technique used by the current implementations of TCP. We wanted the protocol to be robust to ACK loss so that it would provide good performance over networks which exhibit asymmetry, provide consistent throughput during periods of congestion in the reverse path, and for the protocol to yield bandwidth to other sessions sharing the path.

The protocol consists of two major areas of improvement over TCP Reno: congestion control and error recovery. With our error recovery methods we can perform timely and efficient early retransmission of any lost packet, eliminate unnecessary retransmissions for correctly received packets when multiple losses occur within a window of data, and provide RTT estimates during periods of congestion and retransmission, i.e., eliminate the need for Karn's algorithm. The congestion control algorithm allows us to determine when congestion exists in the forward path - a condition which cannot be detected by a round-trip time estimate. This type of monitoring allows us to detect the incipient stages of congestion and to respond to these early warning signs by increasing or decreasing the congestion window. In addition, based upon relative delay calculations, we are able to isolate the forward throughput from any congestion that might be present along the reverse path.

The rest of this section describes the mechanisms in TCP-Santa Cruz used for congestion control and error recovery, in that order.



 
next up previous
Next: Congestion Control Up: Improving TCP Congestion Control Previous: Previous Work
Chris Parsa
2000-01-25