Impatience of Internet users is a major problem for its quality of service. A vast amount of data is wasted on transfers that are aborted before completion.
CWI researchers have developed new models that incorporate this user impatience. As a result they are better equipped to deal with aborted data transfers and can help to improve the quality of service for all users. Researcher Maria Remerova will defend her thesis ‘Fluid Limit Approximations of Stochastic Networks’ on this topic on Tuesday 20 May at the VU University Amsterdam.
When a large number of customers use the same service, for example mobile Internet on the same network, the service rate may drop so low that users become impatient and abort service. This user behavior is a significant problem on the Internet. Opposed to for instance drop-outs in call centers, resources have already been spent to serve such users. Thus, protocols are needed to direct the traffic to guarantee a satisfying quality of service for all customers. Customers who leave the network before their service has been completed are usually not considered in these protocols, but can have a huge impact especially when the system is in overload. In her thesis, Remerova developed new methods that are able to assess the performance of protocols in overloaded systems and system with user impatience.
Developing and improving these results requires deep mathematical understanding of complex networks, and is the subject of continuous research in mathematics. This research was funded by the Netherlands Organisation for Scientific Research (NWO).
Thesis: Fluid Limit Approximations of Stochastic Networks
By: Maria Remerova (CWI Stochastics group)
Promoters: Prof. dr. A.P. Zwart (CWI/VU) and prof. dr. S.G. Foss (Heriot-Watt University)
Date: 20 May 2014, 13.45h
Place: Aula VU, De Boelelaan 1105, Amsterdam
Image: Shutterstock
