Evaluating rate-estimation for a mobility and QoS-aware network architecture

In a nearby future wireless networks will run applications with special QoS requirements. FHMIP is an effective scheme to reduce Mobile IPv6 handover disruption but it does not deal with any other specific QoS requirement. Therefore new traffic management schemes are needed in order to provide QoS guarantees to real-time applications and this implies network mobility optimizations and congestion control support. Traffic management schemes should deal with QoS requirements during handover and should use some resource management strategy in order to achieve this. In this article a new resource management scheme for DiffServ QoS model is proposed, to be used by access routers as an extension to FHMIP micromobility protocol. In order to prevent QoS deterioration, access routers pre-evaluate the impact of accepting all traffic from a mobile node, previous to the handover. This pre-evaluation and post decision on whether or not to accept any, or all, of this new traffic is based on a measurement based admission control procedure. This mobility and QoS-aware network architecture, integrating a simple signaling protocol, a traffic descriptor, and exhibiting adaptive behavior has been implemented and tested using ns-2. All measurements and decisions are based on DiffServ class-of-service aggregations, thus avoiding large flow state information maintenance. Rate estimators are essential mechanisms to the efficiency of this QoS-aware overall architecture. Therefore, in order to be able to choose the rate estimator that better fits this global architecture, two rate estimators - Time Sliding Window (TSW) and Exponential Moving Average (EMA) - have been studied and evaluated by means of ns-2 simulations in QoS-aware wireless mobility scenarios.