quaSI-stationary changes in system states observed by one type of traffic, due to changes in other traffic type s) ,are rare The holding times of long-lived traffic are much longer than those of short-lived traffic Short-lived traffic can approximately reach steady state while connections of long -lived remains d unchange
quasi-stationary • changes in system states observed by one type of traffic, due to changes in other traffic type(s), are rare. ◦ The holding times of long-lived traffic are much longer than those of short-lived traffic. ◦ Short-lived traffic can approximately reach steady state while connections of long-lived remains unchanged. 16
The model a circuit switching network with edge-disjoint alternative paths Long-lived calls · Short- lived calls Preemptive priority of long -lived calls Poisson arrival of connection requests exponentially distributed holding times Mean holding time of long-lived calls much longer than short lived ones (200) a maximum number d of overflow attempts trunk reservation to reserve certain resource to primary path connections
The model • A circuit switching network with edge-disjoint alternative paths • Long-lived calls • Short-lived calls • Preemptive priority of long-lived calls • Poisson arrival of connection requests. • exponentially distributed holding times • Mean holding time of long-lived calls much longer than shortlived ones (200) • a maximum number D of overflow attempts • trunk reservation to reserve certain resource to primary path connections 17
Network blocking probability by efpa Fixed-point iterations Quasi-stationary for short-lived connection Initial values of trunk blocking probability Calculate offered load for each trunk Calculate blocking probability Steady state for each trunk probabilities converge or not? No YES Network blocking probability
Network blocking probability by EFPA 18 Initial values of trunk blocking probability Calculate offered load for each trunk Calculate blocking probability for each trunk Converge or not? Network blocking probability YES No Steady state probabilities Fixed-point iterations Quasi-stationary for short-lived connection
Network blocking probability by oPCa Different trunk blocking probability for calls with different numbers of overflow Calculate the blocking probability layer by layer - Layer o Layer 1 -layer d
Network blocking probability by OPCA 19 • Different trunk blocking probability for calls with different numbers of overflow • Calculate the blocking probability layer by layer – Layer 0 – Layer 1 – Layer D