A queueing network model of TCP performance

• Authors: Debessay Fesehaye Kassa;A. E. Krzesinski
• Venue: SAICSIT '05 Proceedings of the 2005 annual research conference of the South African institute of computer scientists and information technologists on IT research in developing countries
• Year: 2005

Measurement, simulation and analytical models are the techniques and tools that can be used to understand and investigate the Internet and its performance. Measurements become costly and inflexible with the growth and complexity of the Internet. Simulation models do not scale with the growth of network capacities and the number of users. Computationally efficient analytical models are therefore important tools for investigating, designing, dimensioning and planning IP (Internet Protocol) networks.This study presents a simple, fast and detailed analytical model of the Transmission Control Protocol (TCP) which is the dominant transport protocol for the end-to-end control of information transfer. The model gives Internet performance metrics, assuming that only basic network parameters such as the network topology, the number of TCP connections for large file transfers, link capacity, distance between network nodes and router buffer sizes are known.The performance metrics are obtained by using TCP and network sub-models and solving them using a fixed point algorithm. Each of the TCP sub-models is developed using a closed network of ./G/∞ queues where each queue represents a state of a TCP connection. Each network sub-model which represents the output interface of an IP router with a buffer capacity of K -- 1 packets is modeled using an M/M/1/K queue.Numerical results based on comparisons against ns2 simulations show that our model is as accurate, yet simpler and computationally more efficient than another well known TCP model. Our model can therefore be used to rapidly analyze network topologies with several bottlenecks and obtain detailed performance metrics. Our model also gives closed form expressions for important TCP performance values and distributions.