Webcast shows off 10 gigabit Ethernet
HP predicts declining growth

No bad questions, no matter how late

More than six years beyond HP's exit the market announcement, customers are still just starting their migration plans. Up on the Linked In networking site, one member posed this question.

A relative of mine (who doesn't have access to linkedin.com yet) is taking proposals for migration strategies for his HP 3000 machine and all the software and data which resides on the machine. HP has recently quit supporting the HP 3000, so a new machine is required, and something needs to happen to all the COBOL and databases on it.

One bid received will cost 10s of thousands of dollars simply for a migration plan, with the actual migration being a separate bid. This particular migration will be to a Windows environment upon which the same or similar COBOL will continue to run. Any suggestions?

Mark Stoddard
Software Engineer, APT Automation

The suggestions have been around for years now, but since we're Linked In members (you can be one too), we posted this summary reply for Mark's relative who's facing migration's challenge.

Since HP canceled its 3000 business plans in 2001, I've heard from hundreds of smallish businesses faced with this problem. We don't sell services. We interview and publish reports on how customers and service providers solve these problems.

Your relative needs to answer a question to decide how to proceed.

1. What is the primary reason for migrating off the 3000?

A) HP support is no longer available
B) The system cannot keep up with computing needs
C) HP's business strategy no longer includes new HP 3000 models.
D) A mission-critical application provider is quitting on you, its customer

Mark should know

A) More responsive and less costly independent support has been available for many years for the system. HP will support HP 3000s for another two years.

B) A 3000 system that cannot keep up with computing needs can be replaced for many dollars less than a migration plan. If the computer is a 9x8 or 9x7 server, a replacement system is especially economical.

C) HP changed its plans for the 3000 because of the small growth rate of the business -- not because of a flaw in the system. HP 3000s will continue to run through 2027. If purchasing a computer with a long future in front of it is the goal, I wish your relative good luck at finding something with an assured path of improvement. Critical mass determines business futures. Windows has more than any choice today, but it comes with its own hurdles to consider.

D) You've got to move away unless you can acquire and keep up source code for the application. This is the most compelling reason to migrate as soon as possible.

Windows is having its heyday now, Linux is on the rise, and the future of HP's Unix presumes to run through 2016. As HP 3000 owners know, however, things can change overnight. No guarantees.

Any migration plan is going to cost at least tens of thousands of dollars, and the execution and testing will cost even more. Assume that financial burden, or extend the use of the 3000 to a point where the migration budget becomes available, through setting aside part of the IT budget each year for the project. Assign the planning for your first expense, to determine the budget. This is an essential first step.

Migration makes sense once you have budget to spend wisely. Many HP 3000 sites have hurried migrations, at great cost, only to see HP extend its system support.

Another suggestion: Make sure your target for the project includes an improved database, system and operating environment, as well as a broader range of application providers. Migration simply because HP has changed business plans, doesn't make fiscal sense.

Specific solution suggestions: Consider AcuCOBOL for the target compiler, if the applications being migrated are in-house; it's most like the 3000's COBOL II. Assess Eloquence for the database, since it's gotten rave reviews and emulates the work of IMAGE, your current 3000 database. Luckily, both are available on Windows.

As for Windows, we hear that a common mistake is to choose Windows as a migration target on the basis of familiarity with the platform, and low capital acquisition cost. If either of these are the primary reason for picking Windows, think harder. Maintaining Windows is a budget-eating task, and the environment is changing all the time -- so whatever familiarity your relative's IT group might have with Windows is going to shrink, unless there's training and good Windows support in the picture.

There's no lack of good companies to inventory and assess your relative's needs in migration. Tell everyone to pursue the project with care, as well as tune out the "we gotta hurry" you will be told by some companies. Unless your application supplier is turning off 3000 support, the timetable to migrate is yours to choose. Major industry leaders plan to use their HP 3000s beyond 2010. The only entity that has quit supporting the 3000 is Hewlett-Packard's marketing and manufacturing departments. Even HP development resource, albeit pared back significantly, works to this very day for 3000 sites. HP still sells software for the 3000.

In 2001 HP predicted the end of the 3000's ecosystem by the end of 2006. Now the company is proceeding through 2010 with limited support. Judge wisely. You have time to make the right decisions.