A method of scheduling requests from various services to a data storage resource, includes receiving service requests, the service requests including metadata specifying a service ID and a data size of payload data associated with the request, at least some of the service IDs having service throughput metadata specifying a required service throughput associated therewith; arranging the requests into FIFO throttled queues based on the service ID; setting a deadline for processing of a request in a throttled queue, the deadline selected in dependence upon the size of the request and the required service throughput associated therewith; providing a time credit value for each throttled queue, the time credit value including an accumulated value of the time by which a deadline for that queue has been missed; comparing the time credit value of a throttled queue to the time required to service the next request in that throttled queue.