基于服务量差值的Virtual Clock接入允许控制算法

The admission control algorithm based on difference of service for the virtual clock packet scheduling algorithm

VirtualClock算法的接入允许控制算法仅要求系统中复用的各个连接的服务速率总和不超过系统的容量 ,但是这并不足以保证每个分组的时延都达到要求 .主要原因在于 :这种接入算法没有对每个连接占用的带宽在何种条件下能够释放作出规定 .文中提出 ,当一个连接拆除时 ,只有等到所有未被服务完的连接在VirtualClock系统中的服务量超过它们在各自参照服务器中的服务量时 ,该连接占用的带宽才能释放 .采用这种接入允许控制算法 。

The existing admission control scheme for the virtual clock algorithm only requirs that the total bandwidth of each session be less than the system capacity. But it can not guarantee the packet delay. The reason is that it does not prescribe the condition when a sessions band width can be released. In this paper, it is proposed that the band width of a session can be released only when the virtual clock system provides more service for the sessions being served than these sessionss ideal servers. Theoretical and simulation results show this scheme can fully guarantee the delay of packet.