Just to reacquaint myself with it. I never used it heavily but I had a housemate when I was at university in the late 90s who picked up some old VAX and Alpha gear and ran a small cluster in our basement.
In the $DAYJOB I work on a new not-UNIX-derived operating system and almost everyone working on it comes from a UNIXy background. I have vague memories of old non-UNIX operating systems (VMS, Domain/OS, etc) with ideas that could inspire us, but I’ve been mostly using UNIX-like platforms myself so I normally have more hand-waving than concrete suggestions to share.
Rewriting applications that live on OpenVMS now , generally totally sucks(I’ve done it once). Nobody wants to do it. It’s easier and cheaper to just license OpenVMS and continue down the dead end road. Eventually everyone will move off, but it might take a few decades.
Sure, but gradual migration will happen and very few users will do greenfield VMS. Even under HP, VMS was a fourth-place OS by revenue, after HP-UX, Nonstop, and Windows Server for Itanium. It may be supported for a long time, but I can’t imagine it’s going to have a particularly aggressive pace of new feature development.
I consider myself knowledgeable about computers but I’d never heard of Nonstop before. Fascinating! This Wikipedia entry smells very PR-like but can serve as an introduction
I’m no fan of Nonstop - but it definitely has some cool features. It’s basically in a weird midpoint between Stratus VOS and IBM Z in how it’s positioned, and its customers mostly love it. It’s also the only OS that HPE bothered porting off of Itanium (though they did an internal UX port for x86 at one point, which was canned before shipping.)
I agree, but it has to be supported on hardware that exists in the meantime. VMS was DOA without hardware to run on. They HAD to transition to new hardware.
The difference is that Itanium was never a huge market success. VMS was a MASSIVE success. A quick cite I just looked up says that over 400K machines were sold.
I can’t find good numbers on Itanium servers sold, but I’d guess it’s nowhere near 400K.
Respectfully, look outside your bubble. There are more lines of VMS code wriggling around out there in the cold, dark reaches of the computing landscape than many of us realize.
Just because we don’t see it doesn’t mean it’s not there.
VMS was the same kind of success that first generation of IBM PC was. Those machines were built to be INCREDIBLY rugged, like, mil-spec “we shot the power switch 900 times with bullets to ensure it could still operate”^1
1: This may be apocryphal, but was told to me by someone who served on the committed that decided when new commands etc. would become part of VMS.
My bubble? I worked on VMS systems in production from 2010 on, on Alpha and IPF, across multiple generations of hardware. And near-universally, what I heard was “we’ll keep this as long as it ain’t broke, but we’re sure as heck not pushing new workloads onto this.” Last I heard, there were a couple thousand on-contract VMS sites remaining (HP says 2500 as of 2014), of which a fair share are in the process of migrating.
The “not a success” Itanium was moving four billion dollars per year in hardware at its peak, and that peak wasn’t in the 1970s. And VMS was something like 2% of that $4bn.
IPF was successful for HP. It outsold Opteron (by revenue - not by volume) by a significant margin at its peak. It obviously did not become the dominant merchant processor family that Intel was hoping for, but it did well from HP’s point of view.
I wonder if a part of the issue is one of perception.
Most of us by that time were out of the mindset of workstation vendors of old, whereas HP still had plenty of giant gov/enterprise customers to sell to, so plenty of money to be made.
I’ve flagged this as spam: I don’t think we need announcements of an impending announcement. And this is a second such pre-announcement.
Just wondering - do you see this as the most effective way to say “I don’t think this belongs on Lobsters” ?
Because I personally see spam as a qualitatively different kind of animal - like, “GROW HAIR QUICK” kind of deal.
This would be off-topic then?
One day later, the release announcement is here: https://vmssoftware.com/about/news/2022-07-14-openvms-v92-for-x86-announced/
Sadly no download available, just a link to contact sales… Let’s hope the community program is updated soon
I’m really hanging out for a community license.
Just for some fun exploration? Or do you have plans to make something interesting?
Just to reacquaint myself with it. I never used it heavily but I had a housemate when I was at university in the late 90s who picked up some old VAX and Alpha gear and ran a small cluster in our basement.
In the $DAYJOB I work on a new not-UNIX-derived operating system and almost everyone working on it comes from a UNIXy background. I have vague memories of old non-UNIX operating systems (VMS, Domain/OS, etc) with ideas that could inspire us, but I’ve been mostly using UNIX-like platforms myself so I normally have more hand-waving than concrete suggestions to share.
I have a hard time seeing how it survives long-term, but I’d like to be surprised.
Rewriting applications that live on OpenVMS now , generally totally sucks(I’ve done it once). Nobody wants to do it. It’s easier and cheaper to just license OpenVMS and continue down the dead end road. Eventually everyone will move off, but it might take a few decades.
Sure, but gradual migration will happen and very few users will do greenfield VMS. Even under HP, VMS was a fourth-place OS by revenue, after HP-UX, Nonstop, and Windows Server for Itanium. It may be supported for a long time, but I can’t imagine it’s going to have a particularly aggressive pace of new feature development.
I consider myself knowledgeable about computers but I’d never heard of Nonstop before. Fascinating! This Wikipedia entry smells very PR-like but can serve as an introduction
https://en.wikipedia.org/wiki/NonStop_(server_computers)
I’m no fan of Nonstop - but it definitely has some cool features. It’s basically in a weird midpoint between Stratus VOS and IBM Z in how it’s positioned, and its customers mostly love it. It’s also the only OS that HPE bothered porting off of Itanium (though they did an internal UX port for x86 at one point, which was canned before shipping.)
I agree, but it has to be supported on hardware that exists in the meantime. VMS was DOA without hardware to run on. They HAD to transition to new hardware.
The difference is that Itanium was never a huge market success. VMS was a MASSIVE success. A quick cite I just looked up says that over 400K machines were sold.
I can’t find good numbers on Itanium servers sold, but I’d guess it’s nowhere near 400K.
Respectfully, look outside your bubble. There are more lines of VMS code wriggling around out there in the cold, dark reaches of the computing landscape than many of us realize.
Just because we don’t see it doesn’t mean it’s not there.
VMS was the same kind of success that first generation of IBM PC was. Those machines were built to be INCREDIBLY rugged, like, mil-spec “we shot the power switch 900 times with bullets to ensure it could still operate”^1
1: This may be apocryphal, but was told to me by someone who served on the committed that decided when new commands etc. would become part of VMS.
My bubble? I worked on VMS systems in production from 2010 on, on Alpha and IPF, across multiple generations of hardware. And near-universally, what I heard was “we’ll keep this as long as it ain’t broke, but we’re sure as heck not pushing new workloads onto this.” Last I heard, there were a couple thousand on-contract VMS sites remaining (HP says 2500 as of 2014), of which a fair share are in the process of migrating.
The “not a success” Itanium was moving four billion dollars per year in hardware at its peak, and that peak wasn’t in the 1970s. And VMS was something like 2% of that $4bn.
That’s interesting. Thanks for posting. I’d been finding it hard to put a finger on the pulse of Itanium’s success or lack thereof.
And I apologize for suggesting you look outside your bubble when you’re clearly qualified to speak on this topic.
I’m just very used to people posting things like “JAVA IS DEAD! Who uses that anymore?” when it’s anything but :)
IPF was successful for HP. It outsold Opteron (by revenue - not by volume) by a significant margin at its peak. It obviously did not become the dominant merchant processor family that Intel was hoping for, but it did well from HP’s point of view.
I wonder if a part of the issue is one of perception.
Most of us by that time were out of the mindset of workstation vendors of old, whereas HP still had plenty of giant gov/enterprise customers to sell to, so plenty of money to be made.
In a business where we spin up new servers in milliseconds, what’s place of servers that stay up for decades?