Improved Service Curve for Element With Known Transmission Rate

Network calculus is a well established method for computing guaranteed delay bounds in networks. When a network element is shared between several flows, its aggregate capacity is shared with regards to a scheduling policy. In this case, network calculus can compute a residual service curve for each flow, based on scheduling information. In most systems, when a frame is selected for transmission, it is transmitted at full (constant) line rate up to completion or preemption. In this letter, we propose a generic approach that can enhance any residual service curve by taking this effect into account.