Could not download file: This paper is available to authorised users only.


T. Álvarez, L. Nicolas, V. Álvarez

University of Valladolid (SPAIN)
Simulation is a widespread procedure for testing and understanding how real systems evolve. Internet has a great impact in our society. Everyone wants fast and reliable links. But there are situations when congestion happens because the network delay and the number of users fluctuate from one moment to the next.
There are different approaches for dealing with congestion: classics (such as drop tail or RED) or process control techniques (PID, predictive control or fuzzy).
Future Telecommunication and Computer Science engineers need to learn about this topic. They learn networks, Internet and their protocols. So, we need to teach them how TCP deals with congestion.
It seems logical to simulate this problem, but, which tools should we use? Should all the different techniques be implemented? or should the student learn how to program some mechanisms?
The answer is not unique. It depends on the type of engineers and the course level. Usually, Computer Science students prefer to program something in order to understand it. Telecommunication people do not feel so urgently this need.
We have chosen Opnet as simulation tool because it works with state finite machines, so real network events are handled as close as reality as it could be. The possibilities for adding new algorithms are great but not straightforward. A very good knowledge on how the software works is required.
So our strategy is two fold:
- We provide the students with some basic examples and step by step instructions on how to simulate routers, queues and basic congestion control algorithms (drop tail and RED).
- Another tutorial explain what to change and what to include and where if new algorithms are going to be tested. CHOKE and DRED are described and results compared with the first tutorials.
From this point onwards, students can simulate more complex topologies and program new strategies without having to spend quite a lot of time in programming issues.

This paper presents a step by step description of the tutorials and the experiences with students. Results have been encouraging and more algorithms will be added.