SRD Service: A Stochastic Network Calculus Approach for Service Differentiation Analysis

The popularity of delay-sensitive applications has been rapidly evolving in recent years. However, the current Internet is not well designed to support the delivery of such traffic. The SRD (Stochastic Rate-Delay) services proposed in this paper empower a user to choose either a higher transmission rate or a low queuing delay at a congested network link. While average low delay remains guaranteed, the delay constraint is allowed to violate occasionally. The implementation enforces the intended service differentiation via link scheduling and sizing of the buffer. We build a model using stochastic network calculus to analyze the SRD services proposal and shed light on its fundamental characteristics quantitatively.