Call level quality of service management in FCA-based wireless multimedia networks

In wireless cellular networks, the management of hand-off dropping probabilities for different types of traffic is an important call level quality-of-service (QoS) issue. Admission control alleviates congestion in cellular networks and aims at satisfying the required hand-off dropping probability bounds for different types of connections. In this thesis, a new adaptive admission control scheme for fixed channel allocation (FCA) based wireless cellular networks supporting mixed multimedia traffic is proposed. A distributed and event-based call admission algorithm which comprises a technique for estimating future hand-off load and a method for bandwidth reservation in a group of neighboring cells is presented. A metric for determining the overloading of a cell is proposed. It is shown that by monitoring the overload state of the cells and using bandwidth reservation in the admission control policy, the QoS on hand-off dropping probabilities for different types of traffic can be guaranteed. The role of the QoS regulation parameter Pov,T in the call admission process is analyzed for three types of networks: networks with single type of traffic, networks with two types of traffic sharing the capacity, and networks with two types of traffic using partitioned portions of the cell capacity. Techniques for extending the analyses for more than two types of traffic are given. The analytic results are validated by simulations. Comparisons performed show that our scheme outperform the DCAC method [1] in networks with single type of traffic, while partitioning policy may be an efficient yet simple alternative to complete capacity sharing in networks with two types of traffic if the QoS bound on hand-off dropping for narrowband traffic is much smaller than that required by wideband users