Mainframe

Mainframe

In the broadest sense, any large computer—large in the sense of computing capacity, that is, not in physical size--can be classified as a "mainframe" computer. Other classes of computer include supercomputers, minicomputers, and personal computers. Although there is no precise formula for what differentiates a mainframe computer from other types, there are some basic guidelines. Generally speaking, in a multi-user environment a mainframe computer is capable of supporting more users than a minicomputer. Like the typical mainframe computer, a supercomputer is a large and powerful device but, unlike the more general-purpose mainframe, usually serves a few users or only one, and is usually dedicated to solving complex computational problems (such as those found in advanced engineering or scientific applications), rather than to general-purpose use.

The term "mainframe computer" had its genesis in the early days of the commercial computer industry. The major components of the electronic computers produced in the 1950s and early 1960s were mounted on racks or frames. In order to keep the lengths of cables interconnecting a computer's components to a minimum (thereby maximize processing speed), a computer's central processing unit (CPU) and main memory were most often housed together in a single frame, which came to become called the computer's main frame. Other frames of the computer housed peripheral devices, such as the secondary memory.

The first commercial electronic computers were physically huge machines that used vacuum tubes to perform their computations. By the late 1950s, mainframe computers appeared with transistors completely replacing vacuum tubes. Although transistors used considerably less space and power than vacuum tubes, the first all-transistor computers remained large mainframe machines. The advent of the integrated circuit in the 1970s meant that computers could be greatly reduced in size while still possessing adequate processing power for many tasks.

Mainframe computers have often been applied to multi-user and multitasking environments. In a multi-user environment, the mainframe computer can allocate computational resources to multiple users, conversing with each user through a different input/output terminal. Multitasking is a feature provided by a computer's operating system which allows multiple tasks to be performed concurrently.

Unlike personal computers, a mainframe is managed and maintained by professional programmers and system operators. Early mainframe computers were often used in a batch-processing mode with "jobs" being submitted by users using "dumb" terminals (i.e., machines with no built-in logic circuits, capable only of exchanging information with the mainframe computer). The mainframe computer executed the jobs submitted to it and returned the results to the appropriate terminal. As computer technology evolved, mainframe computers became an integral part of distributed processing systems (see distributed systems). In such systems many smaller computers, such as minicomputers, are linked together in a system controlled by a host computer, often a mainframe computer.

Since the early 1980s personal computers have become increasingly powerful . Today, these smaller computers are able to perform many tasks once associated with mainframes. For example, personal computers are capable of performing sophisticated graphics and video-rendering tasks that at one time were solely the domain of mainframe computers or even supercomputers. Moreover, for the cost of a single mainframe an organization can purchase many smaller computers. Software costs, too, are usually higher for mainframes than for personal computers. This is because commercial software packages and tools have been extensively developed for personal computers, whereas the system and application software for mainframe computers is often custom-written. Because of these and other factors, it seemed a foregone conclusion to many industry analysts of the early 1990s that mainframes would soon be completely replaced by smaller, more versatile computers.

However, mainframe systems continue to be utilized for a variety of reasons. In government, military, and corporate environments, mainframe computers are often seen as having several advantages. First, a single mainframe computer may require less maintenance than a network of smaller computers. Another beneficial aspect of mainframes is security and system control. Because mainframe processing and memory storage are centralized, mainframes are often seen as providing a more secure environment than, say, a collection of workstations, each with its own memory and microprocessor. Access to the mainframe can be monitored more effectively than access to a multitude of distributed computers. Finally, organizations are sometimes reluctant to forego their tried-and-true mainframe systems for a new approach (such as multiple workstations) that may have unforeseen impacts on operations.

Mainframe computer systems have been enhanced with new capabilities. Internet software tools, for instance, add new capabilities. Additionally, over the past decade the new disciplines of data warehousing and data mining have provided another useful role well for the mainframe. A data warehouse is a collection of historic data; oftentimes corporations or government agencies archive huge amounts of data for later analysis, or even just on the off chance that it might come in handy some day. To handle such huge data bases requires correspondingly muscular computers. Data mining (sometimes called "knowledge discovery") is a data-processing function designed to uncover new and useful information from collections of data such as those stored in data warehouses.