基于静态流感知的接纳控制研究

Static Flow Awareness-based Admission Control Scheme

维持链路正常负载是业务流性能指标保证的关键,接纳控制是链路负载有效控制方式。相对链路容量有固定业务区分速率界值的静态流感知接纳控制,应用与流拥塞状态相关的判决概率实现业务接入区分。同时,利用优先公平调度机制的隐性测量功能测量链路优先队列队长和链路公平速率值,共同对下游链路流承载状态进行评价,并作为接纳控制进行判决的依据。仿真显示,相对基于测量的接纳控制算法具有更好的链路状态自适应性和较高的链路资源效率。

The maintenance of normal load is a key to guaranteeing the performance index of the service flows. And admission control is an effective method to avoid the overload of the network. Static flow awareness-based admission control, with certain service differentiation threshold, could make a different admission decision for multi-service flows with different decision probability according to the congestion state. The link congestion state, the priority load of streaming flows and the fair rate of elastic flows, could be measured by the priority fair scheduler respectively. And admission decision conditions can be constructed to assess the downlink congestion level. Simulation results show that the proposed algorithm could adapt to the link state dynamically and has a higher cost-effective resource utility than the measurement-based algorithm.