Radio Access Selection in Multi-Radio Access Systems

Future wireless access systems will be characterized by their heterogeneity from technological point of view. It is envisaged that in certain areas end-users will have a choice between various radio accesses (RAs) such as e.g. classical cellular networks (GSM, UMTS, WiMAX, etc), WLAN hot-spots, or over multiple hops via wireless ad-hoc networks. In this scenario it is crucial to select the most appropriate RA for a user at a given location requesting a particular service. This crucial task is performed by the radio access selection (RAS) algorithm. In this paper we evaluate four different RAS algorithms by simulations of a multi-radio access (MRA) system that consists of a cellular UMTS network and a centrally placed WLAN hot-spot. The evaluation compares the performance of the MRA system with respect to the blocking probability, resource utilization, average user throughput and the 90^th user throughput percentile. The simulation results show that combining the resource pools and adding a radio resource consumption or throughput based criteria in the access selection results in almost two times higher supportable traffic load for a blocking probability target of 2%. The average data throughput is increased between 25% and 100%, depending on the traffic load, while the 90% data throughput is increased up to six times.