A Unified ME Algorithm for Arbitrary Open QNMs with Mixed Blocking Mechanisms

The performance modelling and engineering of the next generation Internet, towards convergent multiservice heterogeneous networks, require the use of credible and cost-effective algorithms. In this context, a unified maximum entropy (ME) algorithm is devised for the approximate analysis of an arbitrary open queueing network model with blocking (QNM-B) and bursty traffic, subject to mixed blocking mechanisms (BMs) namely, Blocking After Service (BAS), Blocking Before Service (BBS) and Repetitive Service Blocking with Fixed (RS-FD) and Random (RS-RD) destinations. At each node of the QNM-B, the generalised exponential (GE) type distribution is employed to model the interarrival and service times. Moreover, analytic frameworks are devised for the determination of the first two moments of BM-dependent effective service times and overall interarrival times, towards a node-by-node decomposition of the entire network. The credibility of the ME algorithm against simulation is illustrated by means of typical numerical experiments of an open QNM-B with bursty traffic and mixed BMs.