Slashdot Mirror
Vax, PDP/11, HP3000 and Others Live On In the Cloud
judgecorp writes: Surprisingly, critical applications still rely on old platforms, although legacy hardware is on its last legs. Swiss emulation expert Stromasys is offering emulation in the cloud for old hardware using a tool cheekily named after Charon, the ferryman to the afterlife. Systems covered include the Vax and PDP/11 platforms from Digital Equipment (which was swallowed by Compaq and then HP) as well as Digital's Alpha RISC systems, and HP's HP3000. It also offers Sparc emulation, although Oracle might dispute the need for this.
Take a handful of emulators of very old hardware and open a cloud that host laughably small instances. It will probably work too.
I will probably be asked tomorrow why I've been saying we should consider a roadmap to replace our 15 year old RISC stuff when we could just do this.
What should we work on this year sir? The 15 year old billing system that is mission critical and on unsupported hardware, software, and custom code written by employees long gone or a fifth try at implementing SharePoint that nobody will use?
SharePoint. Got it. Are we going to use consultants paid so well they drive Teslas and Land Rovers again? Let's make sure we don't have clawback for improper billing or properly documented terms or expectations. It is why we're on implementation #5, but you're right, it'll work this time.
My mom says I'm cool.
Up until about the year 2000, I ran a small hardware shop for customers. Gradually, it became clear to me that the value of computers isn't in the hardware, it's in the software and data that they hold.
In response, I reinvented myself and co-developed a company that hosts data for (now) hundreds of clients and tens of thousands of users. Comparing the total hardware value of all our servers to our annual revenue puts hardware expenses (roughly) in petty cash. Servers host a *lot* of data, it's the data and the software used to manage the data that's valuable.
I have no problem with your religion until you decide it's reason to deprive others of the truth.
Alpha is still supported by HP, and OpenVMS on Alpha supported until 2018.
The emulation by Charon of Sparc is 32 bit, not the current 64 bit one. However, you can run 32 bit Sparc code on 64 bit sparc.
There's lots of useful free stuff for people who want to emulate ancient computers at pdp11.org.
Running VAX software just ain't no fun unless you're causing a city-wide brownout with the power drain... :P
I do not fail; I succeed at finding out what does not work.
Having been in a small/medium business consulting realm I have seen many companies go far to long using old technology because "it works". The issue being of course that there is no support from any vendor when something doesn't. Usually my best argument for companies to get off these old systems is that the hardware will certainly fail and spare parts are increasingly difficult to find and expensive. Its great to have an option of emulation of this sort to allow companies to not have to have the huge burden of being forced to use a modern tool with most likely some considerable amount of downtime due to waiting to the last second. On the flip side of that the hidden long term costs of limping by with old, unsupported software will be even more challenging to present to those with the checkbook.
No, VAX is the name of the CPU ISA. VMS is the name of the operating system that was the primary focus of that platform, although you could also get various Unix-class operating systems to run on VAX systems as well (NetBSD and OpenBSD are the main ones today.)
Incorrect.
BSD Unix was born on the PDP-11; the VAX-based Unix OSes started being available in June 1979, whilst the first VAX (VAX-11/780) was released in October 1977, with VMS as the OS. VMUNIX (the Unix OS kernel that supported the VAX's virtual memory capabilities) came out at the end of 1979.
That would be a reasonable thing to do if it bought time.
Year 5? Maybe it's time to hold off on the shiny stuff for a little bit and do the busy work of shoring up the business.
Year 10? Two or three stupid multi-million dollar projects scrapped, but still no work on the billing system? It might be time to reconsider priorities. Or at least consider doing both?
Year 15? I get the feeling that my ability to configure and maintain a resilient system has created a monster. Management assumes it will run forever, and gets to be wined/dined by consulting firms to put up stupid projects.
If we ever finished ANY of the projects we decided to do instead of fixing the old stuff, it would be one thing. But to continue to retry, and fire consultants every year is just wasteful. (no, we're not the government)
My mom says I'm cool.
Apologies. I messed up the link:
v7 Unix.
Incorrect.
BSD Unix was born on the PDP-11; the VAX-based Unix OSes started being available in June 1979, whilst the first VAX (VAX-11/780) was released in October 1977, with VMS as the OS. VMUNIX (the Unix OS kernel that supported the VAX's virtual memory capabilities) came out at the end of 1979.
That is correct. It was based on Bell Labs v7 Unix, which DEC ported to PDP-11 and VAX, and renamed V7M. Ultrix was the follow on to V7M and was first released five years later, in 1984.
Ken Olsen expounded on the DEC's relationship with loved UNIX:
No, no, you're not thinking; you're just being logical. --Niels Bohr
I used BSD Unix on a VAX 11/750 in 1988. It wasn't ready for prime time - to put it nicely. Olsen was absolutely correct. I think a lot of folks who grew up on Linux or the various free BSD spinoffs from the 90s don't appreciate just how far it's come. If you can find an old copy of "The Unix Haters Handbook" browse through it. Some of the stuff is really pedantic, but a lot of it really was problematic.
Unix had a lot of potential, though, and Linux really helped it reach that potential. It's a more coherent system than VMS. On VMS it's pretty much impossible to use piping. That alone made the shell scripts (DCL scripts, that is) many many times longer for the same functionality. The command line tools weren't made to work together. They did have standardized option parsing, though, stuff that Unix never really had but is now handled by libraries.
Up until the mid 90s you could type "rmdir something/" on most Unixes and it would give you the error "something: is a directory". Yeah, no shit sherlock.
The best thing I remember about BSD in those old days was that processes were tied to terminals. And ctrl-z still suspended processes. At the university a lot of folks were used to DOS where ctrl-z exited a program. Of course, if they were in mail and hit ctrl-z it looked like it exited.
So when we logged in the first thing we would do is type "fg". 90+% of the time a mail prompt would come up. We could then !s to get a shell as the mail user. I was never malicious. I would create a login script that would explain that ctrl-z doesn't actually exit and then destroy itself. Fun times.
Do you have ESP?
No, HP has promised no such thing, they've promised mainstream support through 2020 and minimal support through 2025 for the Itanium 8.4 release. They've also announced that VMS Software Inc. is the sole provider for future versions of OpenVMS and VMS Software Inc. has announced intentions to port to x86_64 but they make no promises, and can make no promises on HP's behalf.
There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
then there is what naughty people do, torrent the Z/OS or Z/VM Application Developer CD (ADCD) sets and run the full mainframe monty in Hercules.
A friend told me about that, that's it, a friend. Copyright pirate scum friend. yeah.
What Linux brought to the table was a whole new generation interested in working on Unix. That's big. It's not directly a technological edge but it translates into that.
Not really. A whole generation was already interested as PC's had just become powerful enough to make a full workstation-class UNIX port worth it. BSD got tied up in a big lawsuit and MINIX was a teaching tool. Linux arrived because folks wanted a cheap UNIX clone no one could sue them over. And it was pretty cool for a long time until most distros strayed from being a UNIX clone and adopted BS like systemd that's not even cross-platform..... or even UNIXy.
And I have plenty of idea what UNIX was like before 1990 because I directly used it daily, dialing in from my Atari 8-bit at first. And it was a hell of a lot nicer than most alternatives at the time. NextStep was also a fabulous BSD/Mach hybrid that I still use, they just call it OSX now. You know, the only UNIX variant with a desktop environment that doesn't feel like a perpetual beta release as well as being the only UNIX with a significant amount of the desktop market?
And silly bugs like you mention still exist in modern BSD/Linux distros. It was also probably fixed in a subsequent BSD release. BSD never died or went dormant. Linux never passed BSD from a tech standpoint. In fact, BSD is cleaner and performs better in a lot of scenarios. Even running Linux or SysV binaries.
1BSD for the PDP-11 was simply a few add-ons to AT&T UNIX. 2BSD was really backports of functionality from 4.xBSD. Especially the network stack. The PDP-11 was never a serious platform for BSD development as the CSRG at Berkeley moved to the VAX pretty fast once BSD became a real OS in its own right instead of a few addons for V7. It was THE platform for early AT&T UNIX development though.
If you consider a Pascal compiler and a text editor a true BSD UNIX release, then yeah, you're right ;-)
VMS remained the dominant OS on the VAX platform but most serious UNIX development still happened on the VAX until the later 4.3BSD days. By the time 4.4 hit, RISC workstations were all the rage and the 386 finally became a contender and the minis started dying out. The 4.4 VAX port wasn't even complete if I remember right. Later versions of Ultrix weren't terrible.... it's really dated and a mash-up of 4.2BSD, 4.3BSD and a little SysV but its quite functional. DEC moved to OSF/1 (aka DEC UNIX/Tru64) for the Alpha. There was an early port of OSF/1 for the DECstation MIPS boxes too but I don't know if it ever got released. IIRC, OSF/1 was a Mach/SysV/BSD hybrid.
I still have a VAXserver 3100 in the closet with Ultrix and NetBSD 4 installed. I fire it up from time to time. Slowest 32-bit machine I own but pretty entertaining. Even have an amber screen real VT220 for it. Clocked at around ~11MHz. Capable of about 2.8MIPS.