In this paper we model a communication network as an open queuing network with Mt/Mt/n/m nodes. It is assumed that the mean time between events in the input flow and mean service time are exponentially distributed with rate parameters depending on time t. Each server in a system can fail and be recovered, the parameters of service, failures and recovery also depend on the number of node j and the number of customers in a system. Servers in the network nodes are completely recovered after failures. The recovery time is distributed according to an exponential law. The main characteristics of the performance and reliability of the system are analysed in this paper. We derive the probability that the application received at time t, will be served and the availability of a system.
[1]
Ilya Gertsbakh,et al.
Models of Network Reliability: Analysis, Combinatorics, and Monte Carlo
,
2009
.
[2]
Igor N. Kovalenko,et al.
Introduction to Queuing Theory
,
1989
.
[3]
R. R. P. Jackson,et al.
Elements of Queueing Theory with Applications
,
1962
.
[4]
Martin L. Shooman,et al.
Reliability of Computer Systems and Networks: Fault Tolerance,Analysis,and Design
,
2002
.
[5]
Boris Gnedenko,et al.
Introduction to queueing theory
,
1968
.
[6]
Attahiru Sule Alfa,et al.
Queueing Theory for Telecommunications - Discrete Time Modelling of a Single Node System
,
2010
.