Guest post by “HP has announced the end of support for OpenVMS, the ancient but trustworthy server operating system whose creator went on to build Windows NT. OpenVMS started out as VAX/VMS on Digital Equipment Corporation’s VAX minicomputers, then later was ported to DEC’s fast Alpha RISC chips – before the Compaq acquisition of Digital led to their untimely demise. HP ported the software to the Itanium, but HP isn’t going to bother moving to the last generation of IA64 and support will finally end in 2015.” The article seems to have confused the end of support of VMS (which projected to happen some time after December 2020) with the end of IA64 sales for machines that support VMS and/or the end of support of the Alpha version of the operating system, but it seems either way that the venerable operating system is on its way to meeting the same fate as MPE.
It’s always sad to see an iconic OS die…. But I never liked VMS so I am happy to finally see it go.
“Thank you for making a small command file happy.”
/Sean
It still remember going through VMS’ documentation and being able to find some similarities with the original NT design.
I expect Application running on this will still be running after we are all long gone.
My favorite OS. Ahh well. The writing was on the wall years ago…
True, even before Compaq was bought out by HP the dead end was pretty much announced when a consumer grade company bought out Tandem and Digital then gutted those organisations by turning them into little more than unofficial dumping grounds for Microsoft’s products and services. It is depressing that the old pillars of the IT industry are dying – SGI, SUN, Tandem, Digital being examples in the last decade that have disappeared. it is depressing because the industry before was pluralistic and innovative but all we have now are Microsoft at the centre and OEM’s as little more than unofficial subsidiaries of Microsoft with next to no originality in their hardware/software delivery other than price, what crapware is rammed down the customers throat and the colour of the box the computer is sold in.
This is what happens when you race to the bottom and the only concern is price.
We are seeing this happening in Android ecosystem as well.
And it will happen in any cheap enough OS that gets picked up bye OEM to work as commodity for hardware.
Sadly very few seem willing to invest on OS R&D.
Linux exists, and it’s hard to compete with free. That’s pretty much dooming commercial Unix.
VMS ain’t UNIX
+1, since I cannot vote any longer.
There is this strange idea that everything that is not Windows happens to be UNIX based, as if.
Unix is what killed VMS. Why do you think there was a POSIX layer in later versions of VMS but not a VMS layer like WINE in Unix?
Because at certain point in time POSIX support was a requirement for government contracts in US.
This is the main reason why Microsoft added a POSIX subsystem on NT.
IPv6 adoption is being pushed by the USGv6 program, where the federal government’s new IT gear MUST support IPv6. This has been in place at least since 2010.
What does IPv6 have to do with POSIX?!
computing technology is a commodity now, and that is a very good thing.
Why bother when what all the “reaseach” seems to produce is stinking garbage like Gnome3,Unity, Windows8, and the various Apple Crapple products?
Edited 2013-06-11 19:03 UTC
I know… remember WANG too? That was another decent company that died when the founder died… And I always wished I owned a nice SGI workstation. They were just too expensive to even consider back then.
I know it does no good to look back on what was often, but I do like to think back on the decade or so I spent at DEC. It was a lot of fun.
Wang actually offers a wonderful counterpoint to this decision by HP. Bleeding cash in part from trying to keep their Wang VS line competitive, post-bankruptcy they came to the realization that they *couldn’t* compete. Other companies facing this decision just gave up (e.g. Pr1me). But what Wang did instead was “manage the decline”, putting in a small amount of resources into both hardware and software, providing upgrades for existing customers. Everyone was happy — Wang (now Getronics) has made a lot of money for its small investments, and customers are happy that they’ve not been forced to switch platforms. The last hardware was produced around 2000, and in 2005 Transvirtual produced an emulator that allowed the Wang VS to run as a virtual machine on Linux, which means the VS customers can get steadily faster machines without Getronics needing to pay for any hardware development whatsoever.
In summary, I think this is a bad decision by HP. VMS still has a pretty big install base; there’s no reason HP can’t just milk them for the next umpty years.
It wasn’t all good, hardly anybody could afford those “innovations” of the past days.
I still think VMS was the best OS.
Linux is no alternative to VMS, it is fundamentally different.
Maybe they will open source VMS but I am afraid there are too many patents involved.
http://en.wikipedia.org/wiki/FreeVMS
No activity in three years. I also think that no small part of VMS’ legendary reliability and security are down to the fact that it has been maintained by some pretty bright folks over the space of 35 years, it’s hard to replicated that in an open source project.
Whoah, OpenVMS now has a kill command!
I never especially liked VMS, and I used and developed on it back in the day. I understand why some did, and am sort of sad to see it go, as it represents the decreasing diversity in the OS landscape.
But “single gunshot heard”? Sorry that is just a bit tacky and adds nothing to the conversation.
I would really love it if OSNews stopped reporting rumor as fact, and dialed down the hyperbole…
It’s not a rumour, you can find all the dates in the official HP roadmap. OpenVMS 8.4 on Itanium is the last version, and all extended support will end Dec 31, 2020. OpenVMS on everything else is already dead with extended support ending in a couple of years.
I did not say this was a rumor. I was referring to OSNew’s overal behavior, where they have reported rumor as fact.
HP has pretty much ignored VMS since they acquired it via Compaq so this is deeply saddening but to be expected. They’ve killed through mis-management and lack of advertising and programs to encourage it’s use VMS, Tru64 and in the process killed two of my favorite CPU ISA’s of Alpha and PA-Risc.
The port to IA64 proved that it could be ported to hardware that didn’t require the extensive memory/priv execution that VAX had (read: IA64 hasn’t got the facilities) so it should have been ported to x86, restricted to 1 or 2 processors so people could start hacking around on it. This is how UNIX beats things.. availability to students at universities and colleges. They have it at Uni/Tafe .. they come out into the workforce and are already sufficiently familiar with one tool that will help their work and recommend it strongly.
HP failed here when they didn’t make it available any more for entry level people.
It’s almost criminal.. if you’ve spent any time with the OS coding on it and have learned some of it’s architecture, it really has many strengths you simply don’t find anywhere else.
HP had something great … and they ignored it.
for example?
For example unprecedented stability and clustering. I have worked in large companies with large servers lot of people use all the time, and the sysadmins there thinks that Unix is unstable in comparison with OpenVMS. Linux is unstable compared to Unix. OpenVMS clusters are the best on the planet. You can mix different OS versions, and even different cpu architectures in a cluster, and up to 144 servers. OpenVMS servers measure uptimes in decades. Lastly I heard of a OpenVMS server with uptime of 17 years. OpenVMS clusters never fail, because if a server goes down, the rest are up. And a OpenVMS server never goes down. Some say that single OpenVMS servers are more reliable than IBM Mainframes. OpenVMS clusters are definitely more reliablie than Mainframes. Lot of finance and telco used to run OpenVMS. For instance, when 9/11, the OpenVMS clusters never missed a transaction when the servers in the twin towers went down.
OpenVMS > everything else when we talk about stability and reliability. And OpenVMS clusters are way more powerful. Like Google, that use lot of cheap servers to get high uptime in total. The difference is that single OpenVMS servers have brutal high uptime, and the clusters have even better uptime.
Unix uptime sux in comparison. Linux uptime sux in comparison to Unix. Windows uptime sux in comparison to Linux.
The OpenVMS servers did not have that extreme performance, but extreme RAS instead (reliability, availability and serviciabilty).
There where two big OSes differently architected: Unix vs VMS. Some say that Unix sux in comparison to VMS, when you study them closer. Unix is considered more fancy, and VMS just works. Large finance companies prefers stability to extreme performance, because Enterprise is conservative. When performance is enough, you dont need more. Most servers have enough performance. They just have to work. Like bank systems, they need to be up, no matter what. You never over clock such servers, they are using double and triple hardware to ensure uptime.
Those are not examples, per se, more like folklore/hearsay and personal bias/opinion. I was looking for specifics.
You want specifics on why OpenVMS is better? Like, “OpenVMS can context switch 12.5% faster than an equivalent Unix kernel”, or “OpenVMS has a maximum of 128 processes, whereas Unix has 92”?
Well, I dont see the point of a listing things that OpenVMS does better. It is like “Windows has 32 kernel locks, and Mac OS X has only 16, therefore Windows is better” or “NTFS has 256 character filenames, and ZFS has only 192, therefore NTFS is better” (I made up all examples). Dont you agree that a list of specifics does not tell the whole story?
If you have the chance, talk with sysadmins instead. Preferable those who worked in Enterprise and on Unix servers and on OpenVMS servers. And worked with high availability clustering. Hear their war stories. I think this tells more than a list. Best is of course to work with both OS yourself for a couple of years, but that is not realistic? Second best alternative, talk with people who knows, or read text by people who knows.
HP didn’t kill VMS, the market did. No one wanted it, apart from the US government and some banks. The world belongs to *nix and Windows; that market reality was already well entrenched by the time HP bought Compaq.
While we’re here, Alpha was killed by Compaq, again before HP bought them.
Disclaimer: I work for HP (but am not anything to do with OpenVMS)
I think HP killing off PA-RISC was its big mistake though.
IBM does quite well with AIX/Linux machines running POWER CPUs and HP could be competing in that space with HP-UX/Linux on PA-RISC. Betting on IA64 was a poor choice, but I don’t blame HP for that as Intel of course made the same bet. But at least Intel saw the daylight breaking through the clouds almost a decade ago and re-focused on amd64 *cough* I mean x86-64.
Linux running on POWER servers? I dont see the point of doing that. The old Westmere-EX was only 12% slower than POWER7 on some benchmarks like SAP. Then we had several new generations of x86 and soon IvyBridge-EX will be released which is much faster than Westmere-EX. So why would anyone want to run Linux on an expensive POWER7 server, instead on an cheap x86?
Why don’t you ask them? Because the previous poster wasn’t talking about something theoretical, but something a fair number of IBM customers do run Linux on the Power7 architecture.
I think the reason that HP didn’t go that route was that there are only so many architectures people are going to put up with. And in a field at the top that is x86, Power, and SPARK, there wasn’t room for one more player.
Not to mention the cost involved with keeping the architecture competitive with the previously mentioned trio. It was easier and cheaper to go to x86 (and Itanium) and differentiate other ways.
Maybe you want 32 sockets? Maybe POWER has instructions useful to your problem domain? Maybe its because POWER like Itanium has HA features that haven’t made it into x86 yet?
You make it seem like it is a very “out there” proposition, when it is not. Obviously the common use case is well matched with amd64, and that is the use case amd64 is tailored too.
this happen when HP deciding to put 100% support on GOOGLE ANDROID OS..
KILL MICROSOFT FATHER AND RESTORE PEACE ON DEVELOPERS.
ITS GETTING BAD NEWS FOR MICROSOFT EVERY DAY FROM VISTA WINDOWS 8 THEN XBOX ONE NOW THIS HAPPEN..
WHAT’S WORST CAN HAPPEN TO MICROSOFT NOW..APPLE DECIDING TO OPEN THEIR OS FOR EVERYONE ELSE…THAT MIGHT KILL MICROSOFT COMPLETELY AND CHALLENGED GOOGLE ON PC SPACE..