HP Webcast aims to control transition cost

By John Burke

A cozy crowd of 54 people joined HP in cyberspace on May 22 for the ninth in the ongoing series of HP’s 3000 Transition Webcasts. Subtitled “Controlling Costs,” this Webcast promised to examine the following questions:

What’s the quickest and least expensive way to make the transition? Which tools will save me the most money? And can I save money by simply staying on my HP 3000?

HP’s George Stachnik hosted the Webcast, this time with input from three HP Platinum partners: Birket Foster of MB Foster, Chris Koppe and Nicolas Fortin of Speedware, and Frank Calvillo of MBS. Featured in the Webcast were the Automatic Migration to Unix and Windows (AMXW) tools from Neartek (www.neartek.com). Representing Neartek were Leo McCloskey and Ernesto Soria.

The Webcast focus was on “low cost” solutions, particularly on homesteading and “rapid, automatic” migration using the Neartek tools. But HP first reviewed several items from the recent Solutions Symposium. We’ve already reported on these, but one deserves repeating. While October 31, 2003 still marks the end of sales for HP 3000 systems and chassis, just about everything else will be available for sale after October 31. Note this represents a positive change for HP e3000 customers from HP’s original plans, because it can make homesteading easier, at least until 2006.

While there is no universal agreement as to the best transition strategy, all participants in this and every other Webcast do agree on one thing: planning is the key to success, and lack of planning is the most likely cause of failure. This was emphasized repeatedly during the Webcast.

The Webcast started off with an examination of the pros and cons of emulation mode vs. native mode when doing a migration. In this context, emulation mode means using abstraction layers for TurboIMAGE, MPE intrinsics, and so forth. Generally speaking, emulation mode, compared to native mode, is faster to completion, less expensive initially and requires less training. But it requires ongoing license and maintenance fees, may prevent you from taking advantage of emerging technologies and may just be “postponing the inevitable.”

There was a brief discussion of the “simplest solution,” using Eloquence as the underlying database. However, most of the rest of the Webcast was devoted to the Neartek AMXW solution. AMXW moves the entire HP 3000 environment to Unix, Windows or Linux, utilizing ACUCOBOL or MicroFocus COBOL as the target compiler for HP 3000 COBOL. The emulated environment includes HP 3000 intrinsics, JCL, and screen processing. Databases supported by AMXW include Oracle, IBM’s DB2, Sybase, Informix, SQL Server and Eloquence.

AMXW can be a do-it-yourself migration solution or you can utilize a consulting organization to do it for you. Neartek claimed that typically 98 percent of a migration can be automated. Furthermore, they also said that 4 million lines of code can be migrated by two engineers in two months.

Generally speaking, HP and its Platinum Partners have not been fans of homesteading, even for a few years’ time. This Webcast showed no different spirit from the vendor. The thrust of the message from HP and its partners in this Webcast was that all customers will eventually have to migrate, and it will be more expensive the longer you wait, so do it now.

When HP started the Transition Webcasts, it cautioned against trying to add functionality during a migration or port, because this additional work added complexity (synchronizing the code base, testing concerns, etc.). During the latest Webcast, as well as this spring’s HP e3000 Solutions Symposiums, a different strategy has emerged. In an obvious effort to help justify the costs of a migration, both HP and its Platinum Partners now suggest adding functionality as a way to show some Return On Investment (ROI).

The Q&A session during each Webcast is often the most interesting because it is the only unscripted segment. This Q&A was relatively tame compared to others, but there were still a few surprises. One slide had suggested that in any emulated solution with a TurboIMAGE abstraction layer there would be no need for a separate DBA. This was clarified to mean if you are using Eloquence (the lowest-cost commercial database) you probably do not need a DBA — but if you use Oracle or SQL Server as the underlying database engine, you probably will need a DBA.

Another question was about mixed environments with both 3GL and 4GL code. It turns out the automatic conversion approach runs into problems in this kind of mixed environment particularly in porting the database. However, the Neartek representatives suggested there are ways to handle this situation, although it requires more manual work.

The presentation sloughed over the issue of migrating the screen interface, so several questions addressed this. Apparently the Neartek tools can work with such third party tools as edWin (Ordina-Denkart), LookVP (Cheops) and ScreenJet (ScreenJet Ltd.) as well as with the screen facilities available in the target compiler.

It also came out during the presentation that ongoing support costs for AMWX average around $7,000 a year, but there were no estimates for the product’s initial migration costs. I asked about these initial costs during the Q&A but did not get a clear answer, so I pursued it with McCloskey, the North American AMXW Manager for Neartek. He offered purchase estimates of $25,000 for 100,000 lines of code and $50,000 for 1 million lines of code. AMXW is sold based upon the number of lines of code and the number of target processors.