TELECOMMUNICATIONS
Michael Day
Computers And Communication
Like the brains of animals, computers have evolved from dedicated, single-mindedness into general-purpose information processors. This month's column takes us through a brief survey of the changes in the capabilities of micro-, mini-, and mainframe computers, concluding with how telecommunications are generally used with each type of computer.
When computers were first introduced, they were designed and wired to perform a particular job. When the job was done, either it was performed again with new information inserted, or the computer was rebuilt to run a different job. A patch board reprogrammed the computer by reconfiguring the way the hardware was put together. Such reprogramming was necessary because there was just not enough memory available to hold both the program and data at the same time. (IK of memory was a lot back then.)
Batch Processing
Later, as the available memory size increased (to a whole 4K), software programmable machines were built. This improvement allowed the machine to be automatically "built" by the same mechanism that was used to load the data into the machine. Since most of the time spent on the computer was in "building" it for the particular job at hand, this improvement also permitted an interesting approach to processing, called batch processing.
Batch processing involves loading the program and data into the machine from a mass storage device (usually a tape drive) and running the program. The results are then saved (printed, put back on the tape or on punch cards). The program and data are then purged from the system, and a new program/data job is loaded into the machine. Batch processing helped increase the popularity of these very expensive machines. But they required intensive use to make them worth the cost.
Using The Computer's Time
As computer costs increased, even batch processing was insufficient to offset the costs of the computer. Analysis of computer operations showed that much of the computer's time was spent waiting for information to be given to it. If the computer could be subdivided into individually operating parts (or subprocessors), it would be possible to request the information from a slow external device, such as a tape drive, and while the information was being retrieved, another job could be loaded into the computer and operated on. The processor could later return to the original job and finish it.
Eliminating the computer's inactive or waiting times greatly contributed to the efficient utilization of the computer. By sharing the computer's resources, several different jobs could run at the same time. Careful control of access to the various parts of the system could actually make the computer work like several different computers at the same time. Several users could therefore use the computer without interfering with or being aware of each other. This brought into being time-share computers. Since a user seldom uses the system continuously, someone else could use it when it would otherwise be idle.
Patterns Of Development
When the minicomputer (bigger than a "micro," smaller than a "mainframe") came into being, it went through the same sequence of development. It started out as a computer designed to solve a particular problem and developed into a general purpose machine. The difference: by the time the minicomputer was developed, it was cheaper to design it to be program-controlled rather than to have fixed control. This was true because many parts of the machine could be shared by many parts of the program. Because it was not necessary to have individual parts available for each action the computer performed, the computer could actually "rebuild" itself on the fly.
The result was a shift from the mainframe concept of computing. Since large mainframe computers operate best where there are large chunks of data to be processed, they tend to be run mostly as batch processing machines where an entire job, or a large portion of it, is operated on before moving on to something else. The minicomputer, however, is more suitable to applications where the job requirements are varied and rapidly shifting. They are thus most often found in time-share applications where the ability to handle a large number of jobs simultaneously is more important than the actual processing time. The minicomputer can't meet the raw crunch power of the mainframe, but it surpasses the mainframe in adaptability.
A Rapid Change In Microcomputers
When the microcomputer came along, again the same development pattern was followed. Like the mainframes and minicomputers, the microcomputer was initially developed for single-job applications. But it moved on to more generalized applications more rapidly than either of the other computers. Since the microcomputer was developed as a result of Large Scale Integration (LSI chips), the computers could be created at a very low development cost and an unbelievably low production cost.
The microcomputer too does not have the crunch power of the mainframe, nor does it have the adaptability of the mini. What it does have is low cost of implementation, which makes it the first computer ideally suited to fixed job applications. Some of these applications are found in the calculator, smart thermostats, microwave oven controllers, etc.
In between these fixed applications and the minicomputer are the high level microcomputers (which are coming to be called personal computers). These computers, though sometimes not suitable to the rapidly changing job environment of the mini, do have general processing capabilities. This makes them ideal for personal computing since only a single job generally needs to be run at one time, but the types of jobs that the computer is required to perform are varied.
Telecommunications Needs
You might be wondering, "That's all very fine, but what has this got to do with telecommunications?"
Actually, there is a very definite relationship between the type of computer and its needs in telecommunications. Large mainframes seldom need extensive telecommunications. When they do have such a need, it generally involves special communication circuits designed specifically for the computer system, such as airline or hotel reservation systems, or banking systems. Minicomputers, because they are highly adaptive, tend to use a wide variety of communications capabilities. Examples are the many time-share systems and service bureaus.
Microcomputers, as opposed to personal computers, generally don't have a need for telecommunications. When they do, the telecommunications tend to be specific to the device or application. In fact, in some applications, the microcomputer is the communications device, as it is with some of the high-powered modems available.
Finally, with the personal computer, communications vary depending on the use to which the computer is put. Generally, the application consists of machine to machine communications between users or connection to a large data base service like Micronet or The Source.
These are only generalizations, of course, and it is quite easy to find exceptions to the rule. You can find microcomputers handling multiple communications devices, and fully dedicated minicomputers that have no outside communications at all. As a general rule, however, these basic patterns prevail.