EXAMPLES OF SYSTEM STRUCTURES THAT CAN BE CALLED DISTRIBUTED
A hierarchy of processors is the most familiar data processing structure, with a large, controlling computer at the top of the hierarchy and PCs or terminals at the lowest level. The important characteristic of its structure is that the mainframe, or host computer, is the central, and controlling, component.
Decentralized stand-alone systems do not really form a distributed system at all. They are basically a holdover from the 1960s, when departments put in their own departmental computers, with no intention of connecting them to other systems. Hence, they are decentralized, not distributed.
Systems based on a local area network (LAN) have become widely used as the basis for distributed systems. This approach began in the office system arena with LANs providing the links between PCs, print servers, and gateways to other networks. This structure has no hierarchy, processors communicate on a peer-to-peer basis.
LAN-based systems that communicate with mainframe-based systems another structure for distributed systems. It is essentially a combination of the hierarchy approach (for mainframe-based processing) and the LAN-based system (in offices).
Cooperative systems are the newest member of the distributed system family. A cooperative system melds and extends the previous approaches. All the components are linked to each other via interconnected LANs and wide area networks (WANs). In essence it is an internet-a network of networks. All machines are equal; no machine is at the hub.