A calculus for stochastic QoS analysis

The issue of Quality of Service (QoS) performance analysis in packet-switched net-works has drawn a lot of attention in the networking community. There is a lot ofwork including an elegant theory under the name of network calculus, which focuseson analysis of deterministic worst case QoS performance bounds. In the meantime,researchers have studied stochastic QoS performance for specific schedulers. How-ever, most previous works on deterministic QoS analysis or stochastic QoS analysishave only considered a server that provides deterministic service, i.e. deterministi-cally bounded rate service. Few have considered the behavior of a stochastic serverthat provides input flows with variable rate service, for example wireless links. Inthis paper, we propose a stochastic network calculus to analyze the end-to-endstochastic QoS performance of a system with stochastically bounded input trafficover a series of deterministic and stochastic servers. We also prove that a serverserving an aggregate of flows can be regarded as a stochastic server for individualflows within the aggregate. Based on this, the proposed framework is further appliedto analyze per-flow stochastic QoS performance under aggregate scheduling.