Enhancing Packet Data Performance by dynamic Half-Rate Allocation of Speech Services

As customer demand for wireless data services increases the bandwidth requirements for new applications are rapidly growing Half-rate speech codecs have been introduced in GERAN networks to save operator's spectrum shared by circuit switched voice and packet switched data services. The released physical resources by allocating voice calls on half-rate channels can be efficiently utilized to significantly increase the data rate of packet switched services. Since by nature the speech quality on half-rate channels is inferior to that on full-rate channels a dynamic half-rate assignment strategy has been introduced triggering half-rate allocation by taking into account both the radio conditions of the voice call and the cell traffic load. Provided that the radio conditions are sufficient voice calls are temporarily allocated on half-rate channels during periods of high traffic load to ensure high quality of service level for packet data applications. In this study the trade-off between speech quality and data capacity has been analyzed by means of system level simulations. Exploiting half-rate allocation at the expense of certain speech quality degradation results in a substantial reduction of the number of blocked data calls as well as in a significant increase in the packet data throughput. Yet almost the same performance for packet data services is achieved by pure and dynamic half-rate allocation, the latter strategy provides significant speech quality gain for 50% of the subscribers.