摘要 |
Methods and apparatus are provided for allocating a resource to enqueued requests using a predicted wait time that is based on a prior contribution of the resource to service the requests in a particular queue. A resource is allocated to one of a plurality of requests. Each request is stored in at least one of a plurality of queues, each having a predicted wait time. Once it is determined that the resource has become available, the predicted wait times of the queues are updated based on a prior contribution of the resource to the queues; a performance level of each of the queues relative to one or more service level targets; and the resource is assigned, in response to the determination, to the request based on the evaluation. The service level targets can include one or more thresholds for the predicted wait time. The predicted wait time is based on a prior contribution of the resource to servicing one or more of the queues.
|