Abstract: In response to requests for their use of the resources by the applications, resources are first allocated to “inelastic” requests specified for performance at a specific time slot. A baseline allocation of resources is then determined which is the minimum amount of capacity that has to be used for efficient operation of the resources required to meet the requests. This baseline will include some additional capacity as the minimum efficient capacity is less than the maximum capacity. “Elastic” requests, specified for performance within a predetermined time range comprising two or more of the set of time slots, are then allocated so as to prioritise the use of this spare capacity, bringing actual usage up to at least the baseline allocation of resources. If further capacity is required to meet all the inelastic requests, further capacity is allocated and the baseline redefined to include it.