One Man, One Shop: An Independent View

Michael Anderson is staying busy with HP 3000s this summer, taking steps to make old skills continue to pay off while learning new techniques. Anderson represents the classic HP 3000 shop, one man handling one HP 3000, but during the last year he’s taken a step closer to user groups. Not the national organizations like Interex, but the regional user group that calls itself Greater Houston RUG. Last year GHRUG was one of the first RUGs to invite IBM to its annual conference, a move which Anderson played a part in making happen.

It was not the first time Anderson has taken an independent step in the world of the HP 3000. He arrived at his current place of employment — his organization wants to remain anonymous — in 1999 and helped replace HP’s spotty support with third party options, a move that’s now brought three years of savings and success.

In 2001 he spent less than 25 percent of a budget which had been established to purchase a new HP 3000 — by upgrading the existing HP 3000, resulting in a higher relative performance rating than the brand-new option.

Anderson started out in operations in his first IT job, working his way up from a trade school certificate into programming over the course of a couple of decades, all spent working with HP 3000s. The computer just happened to be at his first job, and so it selected him as much as he chose it. Today he’s utilizing his time-honored skills including FORTRAN/66 and COBOL, but looking forward to learning more about Microsoft’s .NET, and the open-source platforms like Linux — technology far removed from his earliest days when he loaded tapes and pulled 24-hour shifts operating a Series III Classic HP 3000.

We got to meet Anderson while we examined the GHRUG MPE content last fall, and we found an experienced 3000 professional with clear opinions about the future for the computer platform and a willingness to get more involved, even after HP’s decision to leave the system behind. He’s also got an open mind about the role IBM might play for 3000 users who need a big company to turn toward as they turn away from HP. In the weeks leading up to the biggest HP conference of the year, we wanted to talk with a voice that won’t be heard at HP World — but one we believe represents a prototypical 3000 customer, opening the door to what’s next while preserving the value for his organization and himself in what he knows.

What is your organization’s status regarding its HP 3000? Will you homestead for awhile, indefinitely, or not at all?

My recommendation is to homestead for awhile. I don’t see any problem with getting pretty close to 2010. It’s not going to be around forever. Its days are numbered, but that depends on what happens with OpenMPE and an emulator.

What kind of shape did you think your computing platform was in when HP announced its news?

I thought we were in pretty good shape, and I still think we’re in good shape.

In 2001 they gave me a generous budget to get a new HP 3000. I only had to spend about one fourth of it to upgrade the one we had, and that got us a better performance rating out of it that I would have gotten from a new machine. Their Series 979 then carried only the default RAM in it (256Mb), and a single processor. They were using the internal drives, so they only had one IO channel. They said it was slow and needed to buy a new one. I purchased two additional IO channels, three more CPUs, and then maxed it out with RAM. We added two disk arrays, one on each of the new I/O channels. Using the right volume set configuration I balanced our application evenly across the two disk arrays, and made use of the internal disks for temporary, sort, and work files, and this machine screams now.

I was foreseeing that machine taking them out at least another five or six years. We were only using about 10 percent of that machine’s potential for awhile. Now we’re adding things on, like the QSS COBOL Web server [QWEBS]. The HTTP protocol that QWEBS uses does take a lot of resources, but we still have a long way to go before we max out that 3000.

Do you think your relationship with HP has changed since the Nov. 2001 announcement?

It’s hard to answer for my organization. I know the way our IS director looks at it. She sees it from a business point of view. She has a budget to think about, and is

frustrated with HP, their billing and their support. She agrees that the 3000 hardware is extremely reliable, but she’s frustrated with the business side of things.

My opinion of HP slowly started to change in the mid-1990s, when their support for the 3000 started going down. I called for support on the 3000 and someone asked “Is that a printer or a server?” Then they sent HP-UX people out to work on it, and they don’t know MPE. That was the turning point for me.

When I got here in 1999, the folks here were complaining about support, so I went third-party, with Beechglen for OS and software support, and Surety Systems for hardware support. The third party support options are not just cheaper, they are better. These guys don’t need to ask what an HP 3000 is.

So when HP made the announcement in 2001, it didn’t change much for me. I was already on third party support, so it didn’t affect me much. It made me look at the future of the 3000 differently, but I wasn’t bothered by them ending support.

How do you see the future being different?

Just before they made their announcement, the N-Class and A-Class were just hitting the market. They were too new to buy, in my opinion. I was thinking after we went through the five years with our 979, it would be a better system when it was running on an Itanium chip. All the problems would be ironed out with the new architecture.

When they made the announcement that there wouldn’t be any new 3000s, you have to look at something else if you’re talking 10 years down the road. There’s no panic or rush.

I’ve heard a lot of panic, and some are actually in a hurry to jump to HP-UX. I always ask why, why the rush? One time I was asked, “Haven’t you heard that HP announced the end of support for the 3000?” I’d answer “They haven’t been supporting this machine for years anyway. What’s the big deal?” One phase I like to use, that I heard first from [AICS Research president] Wirt Atmar is, “The bits don’t wear out.” He’s thinking you can run this thing another 25 years, and in his environment, he’s probably right. Everyone running MPE needs to keep their eyes open. I’m researching everything I can, and five years from now we’ll have a lot more options available to all of us.