13
I was reading some old questions, and I came across a topic on COBOL where he talked about mainframes. What is a mainframe, and what is the difference between a mainframe and a server as we know it today?
3 answers
8
In general it has always been known as a large computer, as opposed to smaller computers.
They used (or are in some rare cases) to be used for quite complex tasks (by the standards of the time) and run critical applications that need to be centralized. In many cases it was a central computer of an organization that had other computers for other activities.
They were built to take the heat and be more reliable than the rest of our resources. They used to be very modular (even processing could take place in modules, it was as if there were several processors that were different from what we use today). They were even server machines, not as it is today that people use ordinary hardware to serve something by being cheaper and create distributing redundancy (which is often much harder to make work right)
It was very common for them to be accessed by dumb terminals (there was only screen and keyboard that communicated only the characters that entered and left it for the computer without any processing (zero anyway, there was nothing near a processor, although later started to have some little thing and even started to be made by software that emulated the terminal). The name can come from this too, because the terminals were connected to the main board computing, the name that in a certain way today many people call, erroneously, CPU (when they speak of that cabinet that the monitor, keyboard, mouse and other peripherals are connected.
In fact today you have in hand or pocket more computational power than a mainframe had :)
In general it rotates Jobs, which is similar to what many call today’s microservices :D big news! IE, were scripts.
It was common for them to be rented so expensive that they were and required a lot of maintenance. In some cases you would access it for a period of time remotely and pay for the use (know something like this today? :) people think they are using modernities).
Best photo I found to illustrate (there are many around that show the terminal and magnetic tape drive that are peripheral and not the mainframe in itself, there in the background has something, it seems until a perforator of paper tape):
I used one from Unisys :) COBOL even.
In some respects it is a server, but it works a little differently. A little less if you have ever had contact with thin clients that are dumb terminals (although many work on top of an intelligent). It also remembers a little how to access by SSH or RDP.
Server is a broader term and even ambiguous without a context, so it served something yes, but it was not common to be by network as it is today (there were cases of some kind of network, but it was more rare).
2
The main feature of the mainframe is that it lacked microprocessors, in the style of modern computers. The circuits that constituted the CPU (central Processor Unit) were large and modular, occupying entire cabinets. They had advantages, such as redundancy, the possibility of repair, etc.
The ALU (arithmetic and logic unit) was a separate circuit of the CPU itself, which only dispatched instructions.
Old texts still mention the ALU as something separate from the CPU, although every microprocessor (which we call the CPU) has built-in ALU, even the 8 or 4 bit ones.
Nowadays, any centralized computing center with high processing capacity is called "mainframe", but today’s mainframes use microprocessors similar to Unix Pcs and workstations.
-4
Part of the above answer is correct, but they, the Mainframes (with a capital M), are not passed. It imposes respect and can process millions of transactions in seconds and are so fast that a time-Tamp can be duplicated if used as a key, has already occurred to me. You don’t use magnetic tape anymore, I’ve seen it use, and no more punched cards, let alone dumb terminals, and they even use a mouse. Its history is its evolution are fantastic, they are used a lot in banks or large companies to for example make a production planning of a certain product, imagine how you would organize the production of a car, stock, production line, factory floor stock, deliveries, orders to suppliers, etc.... in the banks it overcomes all demand for incoming transactions that must be stored in temporary files (clouds) which undergo updates during the day and in heavy processing at night. ... They still exist, and are giants in. That’s why they’re called High Platform.
Att
-
3
A lot of nonsense in one answer (with all due respect). Magnetic tape is highly used and evolves constantly, Timestamp repetition happens even on "home" machine already MANY years ago, temporary file has nothing to do with cloud, and a lot of other nonsense. Did you really want to give a serious answer, or did you post anything just to see what would happen? (legitimate doubt, no joke)
Browser other questions tagged terminology server cobol
You are not signed in. Login or sign up in order to post.
My first serious contact with Unix was in serial donkey terminal (but it was already modern, SCO Openserver) - Before that there was Gurgel’s Mainframe, it was serial tb. There they used Paradox, probably it was DOS :) - I met some "old things like this" in Banespa, the Hds looked more like toilets, by the shape and size, and had a matrix of 132 columns. I’m glad I got to watch this whole "revolution".
– Bacco
I used SCO but in micro that was Microsoft’s Unix, many people do not know it. It was when I used Fox for database. He could run Unix and Clipper couldn’t. It was a performance beauty because he didn’t have to traffic anything over the network. In one job I made everything terminal (with Harbour) and revolutionized the company. Everything was too slow, and everything became instantaneous. It took 2 hours to make, because I didn’t have all the information from the infra and had to recompile everything to 64 bits and test if it was ok.
– Maniero