As in tethered local networks (LANs), communication protocols (on different protocol layers) and traffic load have a significant impact on the performance of wireless LANs. Additionally, the wireless channel – especially in the context of mobility of the stations – heavily affects the design and performance of WLANs. This project merges these external and internal impact factors into an integrated simulation model. Key submodels encompass Markovian models for autocorrelated traffic generation and autocorrelated packet losses as well as detailed non-Markovian protocol models (IEEE 802.11, TCP/IP) and random waypoint mobility models. The developed models can be validated by measurements with our software monitor with high-resolution timestamping. Similarly, measurements for lowly loaded and traffic-controlled systems serve to determine model parameters (input modeling).
Model building is based on UML-statecharts extended with timing (an implementation of real-time UML) as provided by the commercial tool AnyLogic (by XJ technologies). This tool also allows for discrete-continuous (hybrid) simulation (as required for mobility modeling). The measurement infrastructure has been successfully applied in a web cluster project and been adapted to the WLAN environment. A dedicated graphical user interface provides access to the netfilter framework of Linux and thus a flexible data collection. Several student workers participated in model building and the implementation of the software monitor leading to various student theses.
- 2000-08-01 – 2005-03-31
- “Measurement-based modeling of end-to-end delays in WLANs with ns-2,” Proc. 18th Symposium on Simulation Techniques (ASIM), Erlangen, Germany, pp. 254-259, September 2005 ,
- “Inverse Characterization of Hyperexponential MAP(2)s,” 11th Int. Conf. on Analytical and Stochastic Modelling Techniques and Applications, Magdeburg, Germany, pp. 183-189, Juni 2004 ,