Service based prioritization in (E)GPRS radio interface

General packet radio service (GPRS) and its enhanced version, EGPRS, provide data services with a limited capacity. that is evenly shared by the users. The introduction of prioritization algorithms permits a more intelligent assignment of radio resources, which allows prioritization based on the type of end user service. We use simulations to study capacity improvement in the network by changing the priorities for different services. A weighted round robin queuing algorithm is used for this purpose. Optimal queue weights are calculated which maximize the total amount of traffic carried by the network while satisfying specific quality criteria defined independently for each service. The performance of the network is benchmarked against an identical network with no prioritization. It is shown that the prioritized network can carry more than twice the load compared to the network with no prioritization. Finally, the mutual portions of different end user service volumes are changed to see the effect on optimal queue weights. Using the old queue weights with the new service proportions wastes roughly 10% of the capacity, compared to optimal weights for the new service mix.