In this paper we look at the problem of modelling and analysing an I/O subsystem. There are two conventional ways to look at this problem. The first is to model the I/O subsystem as an M I G I 1 queue, thus neglecting the CPU influence on the arrival stream. The second way is to use the central server model, treating the nonexponentially distributed I/O service times as exponentially distributed.
Here we try tocombine the benefits of both methods. We use a closed network model and analyse this by using an adjusted mean-value scheme. In this scheme service times depend on the number of clients in system. We make use of the residual lifetime distribution of the client in disk service. Some experiments show that the model yields accurate results.