The main competitor isn't other mainframes, but PCs.

A lot of mainframe users never needed a mainframe, they just didn't want to get fired for not buying IBM. Other users need some sort of mainframe reliability, but that's also achievable on a distributed system running on unreliable PCs.

There are some users whose use is genuinely deeply entwined with features mainframes provide, but those are dying out.

> Other users need some sort of mainframe reliability, but that's also achievable on a distributed system running on unreliable PCs.

I don't buy this. Many of our critical systems are on PC architectures. Mainframes don't have some magic sauce, well designed distributed architectures should offer enough reliability.

Edit: reading comprehension fail, please ignore my comment :-)

Even with your edit this is worth replying to:

Why waste your money on a single rack-mounted PC when you can buy 40 cheap cellphones running Android and network them together. That'll probably provide greater reliability.

The obvious answer is the same as why some use mainframes over PCs. You can't easily convert all workloads running on a rack-mounted PC to a network of cellphones. Similarly, you can't easily convert all programs running on mainframes to running on PCs.

Which is what I was alluding to with the last paragraph in my upthread comment. There are mainframe use-cases that are genuinely entwined with those hardware platforms. A mainframe isn't just a fridge-sized PC.

The obvious answer is that a stack of cheap cellphones running Android is a pain to develop for, deploy and manage compared to something actually designed for server use. These are probably not areas where mainframes have the upper hand this century.

You seem to be saying the same thing as the line you quoted: both of you are pointing out that distributed PCs can achieve comparable reliability.

Yeah, you're right, it's too early in the morning for comments. I'll leave my comment up so that others can make fun of it :-p

Back in the day, it was "IBM and the seven dwarfs", Digital Equipment, Control Data, General Electric, RCA, Univac, Burroughs, and Honeywell. Later, Amdahl offered fully-IBM-mainframe compatible systems. Tandem's specialty was fully redundant, fault-tolerant systems.

https://it.toolbox.com/blogs/shayne-nelson/the-60s-ibm-the-s...

By the 1990s, "big iron" referred not only to mainframe (S360 descendants -- now S390 / zOS), but big Unix vendors: Sun (now Oracle), HP, Data General, SGI, etc.

Some of these still survive in some form:

https://en.wikipedia.org/wiki/Mainframe_computer#Current_mar...

I didn't want to explicitly reveal the name, but perhaps I should clarify that it was a software company and it was big relatively to other software companies in mainframe space, not to IBM.

In mainframe hardware, IBM doesn't have much direct competitors. In mainframe software, there are many but most of the are much smaller than IBM, so it is often more symbiotic than competitive relationship.

There used to be quite a few different mainframes on the market. HP/Tandem had their Non-stop systems which was perhaps the most distinct competitor - but there were quite a few companies competing in the space.

However, IBM has dominated this market for as long as I can remember and today I believe they are the only ones making mainframe computers.

Amdahl was notable for having a clone that could run IBM mainframe code.

I think Fujitsu still makes mainframes. That might be it... Hitachi was another one, but they quit the mainframe business in 2017 or so.

Oracle :-|

Fujitsu and Sun/Oracle