|
|
Fig. 1: Terminals are widely used in data warehouses and libraries.
|
 |
Distributed computing also comes in different forms, other than use of the concept on the Internet. Localized non-Internet forms include terminal systems, which have no processor of its own and depends on a central processing server to make it work, and its cousins, the time-sharing system, and thin-client computing (Derfler 45,54). Many libraries employ these efficient yet inexpensive record-keeping terminals, since they are made of essentially one computer hooked to many monitors and keyboards.
Local distributed computing are used in universities, hospitals, and general offices. Computer workstations (please see figure 2) are hooked up to a local-area network (LAN) and can share files with each other and print to one or many printers (Derfler 86-87). Hospitals use distributed computing systems to keep track of patients. All the computers are able to function independently but can share data, which is especially important in keeping databases synchronized.
A different kind of distributed computing shrinks it down to its essence. Parallel-processing machines have one computer with many processors, instead of many computers with many processors. It has a cutting edge in speed over its distributed computing cousin, because the processors are able to "talk" with each other. It can also compute data with variables that depend on each other, such as the previously-mentioned n-body simulation. However, software that makes use of all of these are complex and difficult to program. Ten processors are harder to control than just one. A teacher can control one child more easily than ten children; similarly, ten processors are harder to control than just one.
|
Clustering is in between parallel supercomputers and distributed computing systems. Several computers are placed together in relatively close proximity and connected to each other via interconnectors called nodes, where they can pass data to each other at high speeds. Like distributed computing, it spreads the work across more than one computer. Like parallel processing, the computers can communicate with each other very fast, but still face the same difficulty and complexity in programming the software.
The next section, InsideDC, will introduce you to these forms of distributed computing in-depth and take a look at how distributed computing works from the inside. You will also see the basics of a network, which is the backbone of distributed computing.
|
|
Fig. 2: Computers, such as this workstation, make up local-area networks.
|
|
|
