Adaptive allocation of resources with multiple QoS heterogenous sources in ATM networks - a fuzzy approach

One of the major challenges in ATM based B-ISDN is to guarantee the promised QoS for all the admitted users, while maximising the resource utilization through dynamic resource sharing. In the case of a multimedia call, each of its substreams has its own distinct QoS requirements. The network attempts to deliver the required QoS by allocating an appropriate amount of resources to the admitted calls. Bandwidth allocation play a key role in achieving the promised QoS for every connection or every class of connection. There are dynamic bandwidth allocation polices reported in the literature, but they deal with homogeneous sources viz., video sources only. These dynamic algorithms do not deal with the QoS parameters, and also the main drawback of these schemes is that they lead to unequal queue growth. We present and analyse dynamic bandwidth allocation based on fuzzy logic for heterogenous sources with multiple QoS requirements. We consider that each substream declares a present range of acceptable QoS levels instead of just a single one. As the QoS demand varies, the proposed algorithm allocates the best possible bandwidth to the substream. This is done under the constraint of achieving maximum utilization of resources while simultaneously distributing them fairly among the calls. We validate the proposed allocation method for various heterogenous traffic profiles with various bit rates and QoS parameters and show that the required QoS can be satisfied by appropriately tuning the fuzzy logic controller from the declared QoS parameters and instantaneous traffic arrivals.