BL2xF-channel state-dependent scheduling algorithms for wireless IP networks

Differentiated Quality of service (DQoS) is crucial to the successful uptake of next generation multiservice wireless networks poised to support multimedia traffic with diverse characteristics. One of the key protocols that can be used to provision DQoS in such networks is a traffic scheduler. This article proposes a class of scheduling schemes - best link, lowest/largest "x" first (BL2xF) - that optimizes the usage of the scarce radio resource while considering the QoS requirements of the transported traffic. This is achieved as follows. Applications traffic is classified into QoS classes, each identified wirth a set of QoS metrics and their weights. A scheduling functional relating the QoS metrics of user traffic and the user's instantaneous radio link quality (via data rate) to the serving radio node is formulated. At each scheduling instant, a combination of mobile and QoS traffic class that optimizes the scheduling function is scheduled. An inherent feature in the scheduling scheme is protection of queues of all QoS classes against absolute hoggging. Algorithmic performance is tested via simulated Markov-modeled Rayleigh fading wireless channel.