> Credit card transactions and banking software run on this model for example
Eh, they can but even a couple of decades ago there was a shift to open platforms. 90s and early 00s, sure, it was mainframe and exotic x86 species like Stratus machines. But even then the power of “throw a ton of cheaper Unix at it” was winning.
Banks’ central systems maybe, I have less experience there. IBM did also try for a while to ride the Linux virtualisation wave as well, saying “hey, you can run thousands of Linux instances on a single mainframe”, and I did some work porting IBM software to s390 Linux around 2007.
In the 90s, perhaps not massively, but gaining ground very early in the 00s. I started my career in 2000 and most of the credit-card related stuff I built until ‘05 was targeted at Windows, Linux and Solaris, with a variety of other Unix platforms depending on the client/project.
But the x86 I was referring to in my comment above, Stratus, was (maybe still is?) an exotic attempt to enter the mainframe-reliability space with windows. IIRC it effectively ran two redundant x86 machines in lockstep, keeping them in sync somehow, so that if hardware on one died the other could continue. I have no idea how big their market was, but I know of at least one acquirer/issuer credit card system that ran on that hardware around 2002-3.
Stratus VOS ran on a bunch of non-x86 hardware, i860, PA-RISC, 68000. It wasn't Windows (UNIX admin with a modicum of Stratus VOS experience in production, back in the day).
It seems I encountered the “ftServer” line, which on closer inspection launched in 2001, and was indeed intel/windows 2k, based around Pentium III Xeon Chips.
Meanwhile in 2000 we only considered Linux good enough to host our MP3 file server and quake for the late nights.
All our production stuff was being deployed on Aix, HP-UX, Solaris and Windows NT/2000 Server.
Likewise most of my university degree used DG/UX and Solaris, when Red-Hat Linux was first deployed on the labs, it was after the DG/UX server died, and I was already on the fourth year of a five year degree.
Well we were a small startup, and the idea of using AIX was a non-starter. Solaris was lovely, but our E250 was only for mail, and in hindsight we should have stood up a FreeBSD server with dovecot or something instead of a system that we migrated off of a year later.
We did use NT/2K internally but that was because we had some who insisted on using SMB via Windows.
Such fun times. The nix and nix-like OSes were spreading like fire. I never would have thought I'd ever wrangle them for the majority of my career.
Java was exploding and sun machines were the server platform at the time. Yes, the dot com bubble burst and their stock was in freefall but all the things deployed to sun that survived the bubble didn't just disappear or move to X86 overnight
Just because things hung around didn't mean that Sun/Solaris/Java were long for this world. Linux/x86 was just too cheap compared to SPARC gear. Even if it wasn't as robust as the Sun gear, it just made too much sense especially if you didn't have any legacy baggage.
I never said that was the future, the writing was on the wall to anyone paying attention. All I was saying is that x86 servers weren't nearly as ubiquitos as today
Eh, they can but even a couple of decades ago there was a shift to open platforms. 90s and early 00s, sure, it was mainframe and exotic x86 species like Stratus machines. But even then the power of “throw a ton of cheaper Unix at it” was winning.
Banks’ central systems maybe, I have less experience there. IBM did also try for a while to ride the Linux virtualisation wave as well, saying “hey, you can run thousands of Linux instances on a single mainframe”, and I did some work porting IBM software to s390 Linux around 2007.