April 03, 2012
Leaving legacy: IBM runs Migration Factory
Echoes of the migration bell rung by HP in 2001 are rattling some HP Unix customers loose. During a week when HP's business was considered a carcass by a MarketWatch analyst, news bubbled up on a website in the IBM world about a former 3000-migrating company that's now eliminating HP-UX customers. In a four-month period, hundreds of sites have been converted away from HP's enterprise-grade alternative to the 3000.
That Austin-based consultancy Sector7 has been an expertise resource for migrations since early in the 3000's Transition Era, midway through the previous decade. Some MPE experts have even consulted through Sector7. But even after [some part of] it was acquired by IBM, the company has been able to retain its business motif despite selling its consolidation business to IBM's Global Services group. That portion migrates IBM competitors' systems, whether just a database swap to DB2 or a shift onto IBM's iron in the Power Series. In this world, Unix is considered a legacy carcass.
According to a report at the blog System iNetwork, IBM achieved almost 200 competitive takeouts in the last quarter of 2011 off the HP Unix customer rolls. Each one of these takeouts averages about $1 million per displacement in revenues (although not for Sector7, as we're corrected below by Sector7's Jon Power). That's a yearly total of more than three-quarters of a billion dollars off the backs of HP's Unix, if the analysis from Pund-IT's analyst Charles King is to be believed. From the System i website:
Poaching customers when a competitor is weak is nothing new, and both HP and Oracle have programs to help customers migrate to its own solutions. Still, "I don’t know that any program has worked to the degree that IBM’s has," King says. "IBM is seeing accelerating numbers of migrations both from HP and Oracle. IBM basically has the right tools, and they have a very solid strategy in place to take advantage of uncertainty and concern [in the Unix-focused market]."
IBM and HP have been swiping each others' customers for years, dating back to the days when IBM tried to target HP 3000 shops with the AS/400-Series i systems. There were a few displacements announced around 3000 migrations, but that business didn't show much but exceptions that proved the rule of Windows. However, IBM has had some success selling the System i -- a cousin to the HP 3000 in its integrated design. Some Unix sites have switched to IBM's more proprietary and specialized solution.King's white paper from his analysis house asserts that HP's strategies of Itanium essence and the new Project Odyssey have been helping Sector7 with the displacements. He takes note of the Intel long-term Poulson and Kittson plans, but then says Oracle and Odyssey have been reducing confidence in the lifespan of the Unix legacy.
Continued wrangling between HP and Oracle is doing little to bolster customers' confidence in the platform, In addition, HP's recently announced "Project Odyssey," which aims to "redefine the future of mission critical computing" by developing Superdome 2 systems that support both traditional Itanium servers and Xeon-based c-Class blades, could further confuse the issue.
Every company that's in the business of advising legacy customers strives to portray itself as vendor-agnostic and a trusted partner. IBM even uses that language in describing the Sector7 services. The term "Trusted Advisor" was used in HP's strategic pitches to the 3000 base, before its 2001 migration bell got rung. Both Oracle and HP are serving up those displacements each quarter to IBM's trustees. The scrap between Oracle and HP -- which triggered the Itanium slowdown and Odyssey -- clouds the future for both of these legacy providers.
"Even if HP succeeds or Oracle capitulates," King says, "customers will wonder how deeply or effectively a forced cooperation will extend."
No more trying to figure out what runs on
MPE/iX or where to find it. No more worrying
about availability! www.MPE-OpenSource.org
is all things MPE/iX: Open Source packages,
freeware, scripting, plus loads of tools
and information to keep your 3000 system
alive and thriving!
Thank you for mentioning Sector7. I'll check with my accounts receivable to see exactly where the missing .75 billion dollars went.
I feel I should point out that IBM did NOT buy Sector7. IBM acquired part of our large scale server consolidation business back in 2003.
The complex migrations (Hp3000, OpenVMS, MainFrame, Tandem) are still performed by Sector7, and we still retain many HP3000 experts.
I truly wish that 200 migrations had netted us .75 billion dollars. If it had, I would probably not be writing to your blog.
In all honesty the "free for all" HP 3000 migration spree slowed down about 3 years ago.
OpenVMS migrations have always represented our major market, and number of OpenVMS to Linux requests have increased as geometrically as the HP 3000 have decreased.
Right now, the HP 3000 migration market is in the same place that the PDP-11 was 15 years ago. Fortunately, DEC supplied some kind of route forward with the VAX and Alpha.
Several companies offered replacement PDP-11 boards, then third party maintenance prices shot through the roof. It was glory days at the expense of the users, who either had huge amounts of unportable code (we migrated the 911 FDNY dispatch system -- huge amounts of PDP-11 MACRO and their own real time operating system) or were reliant on special hardware that controlled real time processes.
As with the HP 3000 now, the intensity is dying down, the third party maintenance and manufacturers can no longer make profit from the dwindling numbers, so they are adding to the confusion by abandoning their loyal user base.
Sector7 is almost the "anti-Christ" in the OpenVMS world where OpenVMS is still a religion. However, we and others would not be in business if HP had not abandoned their users by ending the HP 3000, the third party vendors and ISV (the first sign of the sinking ship) had not abandoned their users.
Sure, we're here to make money from these migrations, when most zealots don't want to admit, is that we are the last resort. Everyone else has abandoned the HP 3000 user base and many corporations that have large amounts of custom code are truly worried. Sarbanes Oxley compliance just made matters worse. It was Y2K all over again. That's when most of the HP 3000 migrations were done.
Anyone that is betting their business of revenue from HP 3000 migrations is starting to feel very hungry.
We love our HP 3000 consultants, they are excellent technical resources, honest, hard working, and "old," and yet still love their jobs and work 18 hours a day just because they love the work.
We are in the process of cross training them for the OpenVMS onslaught -- which is now showing all the real indicators of some very late nights and executive platinum status again.
I, like my guys, really have to dig deep for motivation to fly 15 hours for a sales call.
One of the questions that the HP 3000 needs to consider is "Will there actually be someone to turn to in two years?"
I can guarantee that when the HP 3000 migration experts are fully engaged in OpenVMS or other projects it will be difficult to break away.
You mentioned the HP Oracle war; the ISVs announcing end of development is always the beginning. In the case of HP and Oracle, Oracle supplies the relational database used on OpenVMS -- DEC sold it to Oracle many years ago - it's named Oracle Rdb (just to be confusing). It's very fast, very proprietary and very embedded into the coding styles employed on OpenVMS.
One interesting tidbit, like when we were engaged in the PDP-11 end of life -- in the migration business, each migration improves the scope of our tools. When we do the very last migration we have the perfect software migration solution -- and -- no one to sell it to.
As for vendor-neutral -- I can absolutely guarantee that Sector7 does not care what the project is or where it is going. Hp3k, Tandem, OpenVMS, as long as we enjoy the people we are working for, we are happy. Life is too short to deal with clients who technical staff are fighting the decision to move from the legacy platform.
As for the destination -- we have never cared about hardware, whether IBM, Dell, Sun, HP or Oracle. These days everyone wants to go to Intel Linux. We can do the migration work on a RedHat 6 VM running on Mac OS and send it off to run on any hardware without worrying about RedHat 6 hardware issues.
We do care if the client makes a crazy hardware choice, or if the target architecture is going to increase migration cost (Little Endian porting to Big Endian, HP 3000 to AS/400 (Series i) or Mainframe (SeriesZ). Like most other migration vendors we prefer a Unix variant, and these days Linux is definitely the operating system of choice.
Personally, I like the Mac OS. it runs a Unix variant, the hardware it runs on is intelligently limited buy apple which reduces hardware/software incompatibilities, and they make a hell of a fast 16 processor Intel Xenon server, Ok, so it costs a gazillion dollars - but it runs mac os and definitely has the "I'm cool" factor.
Sadly, its likely that they'll discontinue it.
Sorry if this has droned on and on - but - I've not had much time recently for anything but flying and programming, at Sector7 - even the CEO has to get in the trenches and fix bugs.
And - at the end of the day, that's why we love migrations. They can be complex, have tight tim-eframes, and no two migrations are the same. But we do get to meet lots of very cool folks and many different industries who, like me, may be old, and regardless of their operating system religion, we all love technology.
Its been a long time since the guys responsible for programming the HP 3000s have shown anything but happiness and a great willingness to help and be part of the team to move applications -- the programs which have cost hundreds of millions of dollars over the years to develop -- to a new home.
Of course, us old guys remember the reality of development from scratch, being over budget, unappreciated, the long hours, and the corporation looking around for what country can offer the cheapest software engineer, regardless if they've had nothing more than 20 hours of programming experience.
Which is why we love the old HP 3000 guys. An old saying "Measure twice, cut once". in the software world, endless redevelopment is normal. Us old timers" instinctively know how to do the job right and do it right the first time.
That is the key to migrating applications - engineers that just "know" what to do.
Jon Power, CEO
Sector7 (the one and only and original and not part of IBM)
Posted by: Jon Power | Apr 5, 2012 1:54:36 PM
I know the feeling...
Posted by: krikor | Apr 11, 2012 8:29:37 PM
The comments to this entry are closed.