3000's endurance replaced easier than yours
March 27, 2013
System managers who are in charge of HP 3000s might be concerned about the endurance of their hardware. Those who use systems built in the 1990s feel lucky as their 3000 disks keep spinning and the data flows into and out of servers like the Series 929. This is the smallest of the 9x9 3000s, installed in many places as the best 1990s value for entry-level computing.
More than a dozen years later, these 3000s remain on the job. Senior management in these companies might want to ride the lucky tiger as long as they can, to forestall the expense of transitions. However, there's an IT element much tougher to replace than an 18GB drive, a power supply or a processor board.
During an interview this week, a manager who inherited a 929 preached the gospel of newer hardware. It's a problem that has a solution in the wings, as Stromasys makes its way into the homesteading market with its CHARON emulator. This manager said running MPE/iX on Intel PCs sounded "loopy," but he hasn't dismissed HPA/3000. He did look away from a component even more essential than hardware. While that HP iron might go down, the manager going down can also be a major issue. The knowledge of the 3000 is like gold at most homesteading shops, even if management doesn't have a golden budget for the server anymore.
Birket Foster of MB Foster likes to call this the "lottery factor." What if a 3000 manager's circumstances changed overnight, like in winning the lottery? A big annualized jackpot could mean a retirement, and a homesteading company would need a replacement. In-house training before such a change could prepare a company for the day that its 3000 expert goes down, even while the hardware hums along.
This manager's major concern "over anything else, is that I have a super hardware failure, and I can't get any support or replacement parts for my 3000. And while it's down, I'm out of business." Many companies run their HP 3000s around the clock, every day of the week. During the interview, it was suggested that even getting sick could amount to the same concern. That's not in the cards, he answered.
He did have a plan for succession, something a lot of 3000 users haven't formed. The company would hire somebody to come in and learn the 3000 operations over six months, before the IT manager might retire. This can be a difficult situation to engineer as a contingency. If you're not ready to retire, you would find it tough to approach your senior management to say, "let's hire up some IT expertise and make it 3000-ready."
This difficulty becomes a reality at any company where a migration has been "put on the back burner" for 4-5 years, one manager said. Another noted that migration was taking a lot longer than planned, and still another in that confectionary company said migrations have been discussed ever since HP triggered the end of its plans for the 3000. It's money that people are not forced to spend immediately, says Foster. So they don't.
"It's money versus risk where most people end up," he says. "At some point, though, they want to know how much risk they're really facing. It's not really about the hardware risk," he added. In some cases, even a Series 929 could handle twice the business load that it shoulders every day, if sales rocketed. The most critical point of failure is the 3000 expert at the company. Outside help to manage MPE applications, as a backup resource, can mitigate that risk. But it's got to be trained to know your business processes today -- even if senior management sees the 3000 as a less-than-golden resource.
Learning to step in for a manager who goes down, like one at a Florida insurance group did in 2010, takes time. This might be a period where transition planning -- not a migration, but selecting a replacement app -- could mitigate risk over a longer term. The IT pro who knows MPE/iX is the golden goose in these fables.