Issue No.04 - April (1968 vol.17)
N.R. Nielsen , Stanford University
The deliveries of the new third generation computers have presented users with a multitude of problems. Not the least of these is the question of machine configuration. One of the characteristics of the new systems is the flexibility of configuration relative to the older equipment being replaced. No longer can one install "system x." Rather one selects central processor unit (CPU) x and then adds a wide variety of I/O channels, control units, and devices. Equipment selection is not the end of the problem either, for there are many ways to arrange any particular set of devices. Further, not only can configuration changes significantly affect performance, but the effectiveness of any particular configuration is dependent upon an installation's workload. Thus, a configuration which is quite satisfactory for Company Y may be very ineffective for Company Z.
N.R. Nielsen, "R68-12 An Experimental Model of System/ 360", IEEE Transactions on Computers, vol.17, no. 4, pp. 406-407, April 1968, doi:10.1109/TC.1968.226895