论文部分内容阅读
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。
Maintaining the normal load of the link is the key to ensure the performance of the service flow. Admission control is an effective control method of the link load. The static flow admission control with the relative link capacity and the fixed traffic separation rate threshold is applied to differentiate the service access based on the judgment probability related to the flow congestion state. At the same time, the fairness rate of link priority queue leader and link is measured by implicit measurement function of priority fairness scheduling mechanism, which evaluates the status of downstream link flow bearing and acts as the basis of admission control decision. Simulation shows that the relative measurement-based admission control algorithm has better link state adaptability and higher link resource efficiency.