| 摘要 |
PURPOSE: An MLBS(Measured Load-Based Scheduling) method of a router applied to a differential service network is provided to enable a scheduler to dynamically vary a service rate of the scheduler by loads actually assigned to each queue in a node of the differential service network, thereby meeting a service characteristic of a service class and maximizing use of resources of a network. CONSTITUTION: Queues are initialized(S501). If data are inputted to the queues, traffic loads in each queue are calculated at the interval of unit time(U)(S502). Service rates of each queue in a next time section are calculated at the interval of update time(T)(S503). At this time, sum of the service rates in each service class is adjusted not to exceed link bandwidth. For an optional service class(i)(S504), whether data exist in a corresponding queue is checked(S505). If packets do not exist in the corresponding queue and the size of the queue is not larger than 0(zero), the step S504 is returned to pass on service authority to a next queue. If packets exist in the queue of the corresponding service class(i) and the size of the queue is larger than 0(zero), size and processing time(Pu) of the very first packet in the corresponding queue are calculated(S507). The calculated value(Pu) is added to a total processing time(S508). Whether the added value is not more than a preset calculated service rate(Si) is checked(S509). If so, the packets of the service class(i) are dequeued and the step S506 is returned, and if not, the step S504 is returned.
|
