ABSTRACT
In its simplest form, a computer network consists of two or more computers that are able to transmit information back and forth over a communications link. A heirarchical network is one in which a host (supervisory) central processing unit (CPU) and one or more peripheral or local computers are interconnected as shown in Figure 8.2. In a distributed system, the local CPUs are capable of operating for an extended length of time without the necessity of communication with the host system. The length of time which the system can function in an autonomous fashion depends on the specific system design but it is generally related to the changing environmental conditions that the isolated system must accommodate. As long as things are relatively constant the local CPU may be able to operate indefinitely (assuming adequate memory resources), but as conditions change and new input is required from higher level sources, then access to the host becomes the limiting factor. However, alternative host sources may exist in some networks as well as the option of utilizing a manual method of information transfer. While available manual data transfer techniques are less convenient than down loading the data from a host system, in an emergency, these methods of data transmission provide a backup capability that can be utilized in the event that host communications are interrupted. An example is the ability to load a needed program from a backup tape memory so that the local computer system is totally independent of the host when neCeSSalj. For more complex systems, the processor that acts as a host to one system also may be operating under another host. This type of system configuration could have a shop-floor host that communicates downward with a number of shop-floor processors and upward with an area host, and so on.
