Researchers: Jukka Heikkonen, Timo Koskela, Jouko Lampinen, and Kimmo Kaski
Project seeks for tools and models for classifying and modelling the characteristics of the Internet services from the user point of view. In order to achieve this, quality of service (QoS) must be defined in different usage situations. Figure 10 shows the elements of the Internet service in a typical case. Perceived QoS is affected in each of the elements: the user itself, ISP, proxy cache, Internet and the server connected. Each oelement also has a different set of QoS measures, which can be dependent on each other. For example in the network element QoS can be measured e.g. as network throughput, latency time, transfer delay or percentage of dropped data packets. For other elements QoS may not be as easily measured, and consequently statistical models are used instead.
Data for the statistical modelling is collected by using a client installed in the user's computer. Client measures selected Internet services continuously. Data from clients is collected to a database, from which the statistics for certain service can be viewed. For practical reasons the number of different ISP's and individual Internet services measured is limited in the study. Results from a questionnaire study are used to select the most popular Internet services to be measured. A model for the user is built by using statistics of user sessions in the ISP's modem pool and the information from the questionnaire study. Finally different elements and models of their QoS are brought together in order to achieve a specification, which states how to perform the measurements in order to be able to predict the perceived QoS in each case. Results can be validated by comparing the predictions to the perceived QoS.