VMS is alive, well -- and kicking|
By David Berlind, Tech Update
October 19, 2001 2:02 PM PT
Move over IBM, Sun, and Microsoft. When it comes to operating systems that are scalable, reliable, and fully outfitted to run enterprise applications and e-businesses, there's a trusted name with a proven track record. It isn't Unix, Linux, IBM's z/OS, or Windows. It's the venerable OpenVMS, and it's on a path to become more affordable.
Many of Tech Update's readers cried foul when I wrote OpenVMS' premature obituary (see "Fiorina faces competitors, customers, hurdles"). I'm now convinced that the OS shouldn't be overlooked as you forge ahead into the 21st century. In that column, which detailed merger plans of bedfellows HP and Compaq, I quoted Gartner analyst Paul McGuckin, who said that for current customers of both companies, "All roads lead to HP-UX or Windows XP on Itanium." Throw in OpenVMS.
According to the marketing director of Compaq's OpenVMS Systems Group, Bob Blatz, there are currently over 400,000 systems running OpenVMS, supporting over 10 million users. That's nothing to sneeze at, nor is it a business that Compaq or HP should be looking to phase out anytime soon. Indeed they won't. In fact, a detailed road map, stretching into the foreseeable future, addresses the concerns of current customers while keeping OpenVMS in the horse race against other highly scalable and available big iron front-runners.
The main concern regarding OpenVMS' future has to do with the future of the only 64-bit processor it supports -- the Alpha. Alpha's roots can be traced back to Digital Equipment Corp., which was acquired by Compaq. All of the company's Alpha assets were sold to Intel, resulting in the transfer of all Alpha patents, intellectual property, and 200 engineers. Intel was on its own 64-bit path with IA-64, the first fruit of which is the Itanium chip, and has since decided to phase out Alpha, creating some uncertainty for OSes like OpenVMS that run on it.
But, as Compaq's OpenVMS vice president Mark Gorham explained, the company's commitment to the operating system has never been stronger. Officially, Compaq supports its products for a period of five years beyond retirement. But given the OpenVMS group's track record, customers can expect support to go far longer. The company is still supporting 1980's generation VAX systems (running VMS) for many of its worldwide customers. OpenVMS currently supports the EV67 and EV68 generations of Alpha, and will continue to do so moving forward into 2006 and 2007 when Alpha's final iterations EV7 and EV79 will be retired. If history is any indicator, Compaq likely will continue to support OpenVMS on Alpha until at least 2020.
Although porting VMS from the 32-bit VAX to the 64-bit Alpha took hundreds of engineers, Gorham says "all of the heavy-lifting to get to OpenVMS 64-bit architecture is finished." With that big job behind him, Gorham needs only 20-30 engineers to consolidate OpenVMS into a single source pool that, starting in 2003, will support the roadmaps for Alpha (EV7,and EV79) and Intel's IA-64 (Itanium, McKinley, and Madison). This is where some of HP CEO Carly Fiorina's "unifying architectures" rubber meets the road. Fiorina sees IA-64 as the main unifying architecture. Strategically, all HP/Compaq OSes will support it, including the highly fault-tolerant NSK operating system, according to Gorham.
But if all goes as planned, OpenVMS' IA-64 support (code-named "Ruby") won't come at the expense of a smooth transition path for current customers. There's significant overlap between OpenVMS' Alpha and IA-64 lifecycles, and some of OpenVMS' best tricks are yet to come on the Alpha platform. Somewhere in the 2004 timeframe, Compaq will ship 64-processor and perhaps 128-processor Alpha systems based on the EV7 processor family. Code-named "Marvel," these are, Gorham says, "big honkin' systems with lots of memory" that are very scalable, very modular, and very fast, with memory access speeds 300 to 500 times faster than today's offerings.
The overlap also gives an estimated 3,000 developers -- including software giants like Oracle, BEA, IONA, and CompuWare -- time to port their applications and customers. For most ISVs, says Gorham, this process won't involve much more than recompiling the source code for the 5,000 or so OpenVMS applications out there.
Ultimately, the move to Intel's platforms will address three issues that will secure OpenVMS' future. First, the price should drop since greater economies of scale can be achieved with Intel than with Alpha. As a result of that drop, current OpenVMS customers can expect a lower TCO over the lifetime of systems that range in cost from $5,000 to one million dollars. The lower cost will also make the platform more attractive to developers (a key to the long-term success of any operating system). Finally, after lowering the cost and hopefully attracting more developers, OpenVMS could become a more viable option in situations where it would not have been considered before. Gorham says OpenVMS will be running on notebook computers.
But one question remains for many current Compaq and HP shops. HP and Compaq were traditionally rivals in the mini/mainframe market. If the future is so rosy for OpenVMS, what does that mean for the HP OSes (HP-UX and MPE) that OpenVMS traditionally bumped into when competing for a customer? According to Gorham, the only one of the verticals where HP really gave him a run for his money was the manufacturing sector. "There," says Gorham, "will have to be a bake-off based on customer loyalty and the best solution." Citing the pre-merger quiet period that covers both companies, he declined to elaborate. The other verticals served by OpenVMS -- healthcare, government, financial, and telecommunications -- should continue to be served by OpenVMS.
Intel is clearly beginning to level the playing field for almost all OS contenders.
It remains to be seen whether OpenVMS will become a serious option as more prevalent platforms like Unix (including Linux) and Windows continue to gain on OpenVMS' key differentiators of scalability and reliability. But, based on what I've heard, I have to admit: It can't hurt to hear Mark Gorham out before you make your next big platform decision.
Got a great tip? An industry rumor? Or do you want to submit your own column to ZDNet TechUpdate? Send David your submission, and if we use it, you'll be compensated with some of the cool vendor schwag that arrives in our mailboxes on a daily basis.