Modelling of Wireless Networks

OData support
Dr. Lencse Gábor Sándor
Department of Networked Systems and Services

Nowadays almost every technical equipment uses networks to communicate with each other. An average person is probably not aware of the path what an email would take until it reaches the target, when they send an email via their mobile phone.

We are using wireless networks at many fields, like mobile networks, military communication networks, networks using to track motions in consumer electronics or even private networks at home. It is not necessary to install these networks to observe their behaviours, there are suitable simulation software for those purposes. These are called network simulation software. We are able to examine the behaviours of the networks, depends on which equipment or services are available in the network as well as we can change the conditions to see the reaction of the network.

One of the goals of my thesis is to explore these simulators and characterize them by their abilities. I set up a criteria system: the software project has to be live (the source code is maintained) or it is currently used for research purposes; it has to be open-source software; it is an advantage if the source code is available and it can be expanded. In the first phase of my work, I gathered ten simulators which seemed to be suitable for analysing wireless networks and I examined them according to the criteria. I have found four simulators to be suitable for further analysis: NS-2, NS-3, OMNeT++ and the Riverbed Modeler. For further tests, I have built a model of a simple wireless network (with wired components) in these simulators and I created the real network as well as a reference. This model contained two access points and four clients. I examined the simultaneous downloads of files with different sizes to the clients on the real network and on the simulation models as well. The download speeds were more or less close to the test results of the real network but I experienced bigger differences in the execution time results.

All four simulators fit for educational and research purposes because all of them are well documented and they are able to analyse a wide area of network protocols. I have found the Riverbed Modeler and the OMNeT++ outstanding compared to the other simulators.


Please sign in to download the files of this thesis.