=[ x25-box.com ]=

centralized buying policy could easily find itself with a collection of
different machines none of which were being run efficiently.

By the mid-eighties the users had won, mainly because of the technology
revolution that had been going on largely in the background. The
departmental mini then had as much power and access potential as the
mainframe that the DPM was fighting to preserve.

As this battle raged a new issue came to the fore, that of networking.
With the mainframe, as its number of terminals grew, there was an obvious
and consistent network architecture, something like that in Fig. 1.1. The
mainframe always has a star configuration. If terminals are added they
are always a new ray from the centre whether they are connected directly,
multiplexed, or via modems. The strategy is simple and easy to budget.

In the late seventies the same strategy applied to the minis in the
organization. They were all treated as little mainframes and the network
diagram changed to that shown in Fig. 1.2.

By the early eighties this was starting to creak a little. By this time
users had expectations of technology and were starting to think for
themselves about what they wanted to do. Users on their departmental
minis could see advantages in being able to access the machine in another
department as well. Perhaps because they had a software package to

Fig 1.1 Star networks based around mainframe computers

НАЗАД