In partial fulfillment of the Requirements for the Degree of
Doctor of Philosophy

Tsung-i (Mark) Huang
will present his dissertation

Fast Pattern-Based Throughput Prediction for TCP Bulk Transfers

Abstract
The ability to quickly predict the throughput of a TCP transfer between a client and a server, or between peers, has wide application in scientific computing and commercial computing. This dissertation presents a new approach to fast prediction of overall throughput of a large TCP file transfer. The method constructs the time series of windows of segments arriving at the receiver, and predicts future throughput by exploiting knowledge of how TCP manages transfer window size. When the file transfer time series resembles a known TCP flow pattern, this information is utilized for prediction, otherwise simple heuristics are used. Each prediction is also assigned a quality ranking which indicates the expected accuracy of the prediction. We have compared this TCP Pattern-based prediction method against traditional methods like a simple moving average, exponential weighted moving average, and past measured throughput. Experiments were conducted on a large suite of real life TCP traces. Our results show that the TCP pattern-based prediction method performs as well or better than other prediction methods in almost all scenarios. Also, the assigned quality of the prediction is shown to be a good indicator of the accuracy of the prediction.

Date: Monday, April 17, 2006
Time: 1:30 PM
Place: 646 PGH

Faculty, students, and the general public are invited.
Advisor: Dr. Jaspal Subhlok