Blocking Probability in Pipeline Forwarding Networks

As multimedia communications continue to grow steadily on the Internet, pipeline forwarding (PF) of packets provides a scalable solution for engineering delay-sensitive traffic while guaranteeing deterministic Quality of Service (QoS) with high resource utilization. In PF networks resource reservation, while ensuring deterministic QoS on a per-flow basis, can result in a not null blocking probability. A reservation request may fails due to enough resources being available but not during the proper time frames. This work analyses blocking probability of reservation requests since it affects the capability of utilizing network resources to carry traffic with deterministic QoS. The blocking probability and, consequently, the achievable network utilization are analytically derived on general topology PF networks as function of the traffic intensity given the traffic matrix and the network routing. The correctness of the blocking models is also assessed by simulation in different scenarios. This work represent a valuable contribution over previous analytical models of the blocking probability as their application to real size scenarios is impractical due to their computation complexity.