Keeping the MPE Faith, Euro-style

Christian Lheureux first got involved with the HP 3000 and MPE as a student, and more than 20 years later he’s teaching the system’s future to customers in Europe. The director of the OpenMPE organization works for HP business partner APPIC in France, in the heart of the HP 3000 European community. He brings several unique viewpoints to the board of the organization that is dedicated to extending MPE’s life: 14 years of service inside of HP, as well as a European channel partner’s perspective on open systems and the 3000.

Lheureux started working with the HP 3000 on a Series III at business school, then joined HP in 1984, first as a Systems Engineer, then as Account Manager in charge of MPE customers. He worked inside HP’s Response Center, in charge of the center’s late-beta testing of what later became MPE/XL. His responsibilities included delivering performance consulting, assisting migrations, customer and internal training, writing articles, and working with the CSY Labs to debug MPE at source-code level and write patches. Lheureux said he’s done “about everything MPE-related within HP, except manager and salesperson.”

After leaving HP in 1996 to head the High Availability Competency Center in Boblingen, Germany, he then joined Cap Gemini as an MPE Consultant. He started the Systems and Networks Department from scratch at APPIC R.H. in 1998, a business unit that includes the rest of HP’s platforms in addition to the HP 3000. We wanted to ask Lheureux about the 3000’s future options and how the European community is reacting to the transition. We exchanged e-mail in late March, about a month after HP released its first draft of the new MPE license proposal.

What’s your personal reaction to HP’s first draft of the MPE licensing offer for emulators? Is this first offer enough to get the emulation companies working?

The fact that most potential emulator companies I am aware of have started to think in terms of market assessment, price assessment and so forth makes me think that yes, the draft license agreement is a very significant step in the right direction. I am aware of people seeing the bottle half-empty. I tend to now see it more than half-full, and still filling.

Why did you decide to serve on the OpenMPE board?

First of all, I wanted to keep one more option available to my customers. Some of them will leave MPE entirely, migrating, rewriting, replacing or retiring applications. But some of them will want to homestead for some time, possibly forever, for various reasons. Perhaps they can’t migrate within the HP-announced timeframe. Perhaps their budgetary constraints hamper their opportunities to migrate. Perhaps they simply need more time. For whatever reason, there is a segment of our customer base that simply needs one more option. OpenMPE represents that option, hence my decision to apply for a board position.

Besides, for having been a member of the MPE community for more than two decades, I did not want to leave the community without yet another contribution. In a nutshell, I did not want to leave the work unfinished.

Do you believe that OpenMPE should lend its support to a single company’s emulator project? It seems that funding the R&D might be someplace OpenMPE could help.

This decision is up to the Board of Directors, upon advice from the community of OpenMPE members. So far, no single company has been endorsed. Whether this position of neutrality changes in the future remains to be decided.

As far as funding is concerned, I find it hard to imagine that OpenMPE would fund more than one emulator project, if any. So funding one in particular raises at least three prerequisites :

1) Endorse one project, which reverses the Board’s current position of neutrality,

2) Make a technical assessment of which project to fund, most likely based on technical terms, time-to-market, and other terms. This may be the hard part, for we are trying to guess far into the future, and to assess what are still ideas, blueprints, concepts, plans, etc.

3) Provide the funding, which requires to have a funding of our own. So far the memberships are free. There has been a lot of discussion on that subject in the OpenMPE Community recently.

Your AIM nickname is “MPE Evangelist.” Why do you continue to advocate using an operating environment which is being abandoned by its creator?

Not entirely unlike the Apostles went on preaching Christianity long after Jesus Christ had been crucified, I would continue to preach in favor of MPE even after HP no longer endorses it. MPE ought to be adopted into another home. The only reason why a customer would use it or not should be a business decision. If MPE serves customer needs, then customers should have the opportunity to use it for as long as they want. And, as I already said above, not all customers have the time, possibility and budget to migrate. Some of them need to homestead.

What do you see as the practical, hard deadline for leaving the HP 3000 platform? Is this a different deadline than leaving MPE-IMAGE?

The general idea is to “de-couple” the HPe3000 platform and its operating system. The HPe3000 is reaching its end-of-sales in a few months. It will be maintained for another 3 years after that, but will ultimately disappear, little by little. However, MPE needs to live on, were it only for the reason that customers will still need to run MPE-based applications for some time after 2006. Therefore, we need to separate the HP3000, which is destined to die, and MPE, which we expect to live. Hence the proposition of an emulator, which can enable MPE on a totally different platform. The IT market very clearly points toward commodity hardware. Therefore, the decision to advocate an MPE emulation of some form on a commodity platform was a relatively easy and fast one.

So, to sum it up, when we get the emulator, leaving the HP 3000 will not be a great leap forward, because MPE and IMAGE will still be there.

How has the HP 3000 community in Europe been reacting to HP’s November 2001 announcement? Do you hear of many stories of customers already taking steps to migrate away?

To put it mildly, the first few customer reactions have been quite hectic when HP executives were available. The level of frustration has since receded a little, and left the place to more carefully thought-out mental processes.

At this point, though, customers are assessing their options and designing their plans. I hear very, very few stories of actually ongoing migrations. It seems that the option of entirely migrating an MPE environment to, say, HP-UX, does not attract too many sites. Perhaps it is not that appealing to run an MPE environment on another platform, in some kind of compatibility mode.

So what I see is customers taking an application-based approach. They tend to migrate applications (or parts of applications) to another application. This is consistent with another market trend that points toward replacing homegrown or specific applications with ERPs or other software packages. This is another business decision: sacrifice some level of specificity for the sake of cost, thus enhancing the bottom line for the shareholder.

This is also consistent with another deep trend: decision-making processes have been drastically altered over the last few years. Major IT projects, like deploying an ERP, are no longer decided at IT manager/CIO level, but at least one step above, in some cases at CEO level. The CIO or IT manager’s role is to implement the project, not to decide it. So decisions tend to be made on many criteria, among which the technical aspect may only play a minor role. This is comparable to a crew in an airliner: the captain chooses where to go, the mechanical engineer makes sure the airliner flies.

Has the economic slowdown had an impact on HP 3000 customers’ decisions in Europe?

So far, Europe has cushioned the economic downturn a little better than the USA, which is usually what happens in economic downturns. Europe would experience less severe downturns, but would then expand slower after recovery. So we may not be exactly double-dipping, though there are indications of a worsening situation. Still, the current uncertainty around the Middle East conflict may severely hamper or delay a possible economic recovery.

At this moment, nothing much happens in Europe, as far as major IT projects are concerned. It seems that everyone is waiting to see what everyone else is going to do. This is a recipe for a stalled IT environment. As a consequence, of course, HP 3000 migration projects are stalled. And an HP 3000, at this moment, is almost impossible to sell.

HP is ending its business opportunity to sell the HP 3000. Is there still a good business in serving these customers with MPE solutions?

As long as there is a customer base for MPE (notice that I’m mentioning the OS by its name, not the platform it currently runs on), there is a need for customers to be able to obtain service, and there is a business opportunity for actors belonging to the Community. Of course, MPE is probably not the area where we invest most of our [company’s] resources, but, as long as we retain the ability to serve customers, we will stay in that business. We may even have an opportunity to address typical end-of-life management needs: customers operate on scarce human resources, so they tend to refocus their HR toward their main projects, like ERP implementation, and outsource day-to-day activities (which MPE clearly belongs to). We see that as a business opportunity, and we are addressing it.

What’s the most significant change you have observed in HP since its merger with Compaq?

I am seeing two major shifts. The product offering, while still not yet carved in stone, is now complete, with a much more robust Wintel server line, and a few holes plugged in the storage area. One instance is the former gaping hole that used to exist between the VA and XP lines which is now occupied by the EVA line.

Secondly, HPs’ notoriously lousy marketing (no offense meant to my friends within HP, but they are much better engineers than marketing people) is gradually giving way to a much better customer-oriented approach, whether at grass-roots or at decision-maker level. This looks like the continental drift toward smart marketing I did not even dare expect before the merger.

Can HP expect to keep growing its HP-UX business using a reseller-partner sales model, versus direct contact with its customers? Is its Unix alternative good enough to keep customers from choosing a lower-cost environment like Linux?

This is like two questions in one. The reseller-based approach has worked wonders for companies like the one I now work for. We are in the process of becoming a major player in the HP-UX arena, at least in France. Of course, that would never have been possible if HP had decided to stick to a direct-sales model like they had until the mid-80s. At this moment, we are implementing a two-tiered approach, with a lot of involvement by the HP Sales Teams.

Another side effect of the merger is to have much stronger HP sales teams to assist us. This is helpful to give HP more visibility in the field. And, as a single-brand dealer, it greatly enhances and gives a lot of credibility to our HP-only approach. So this is good for business.

Linux is another question. At this moment, it simply cannot do all that HP-UX does. If we go back, say, a decade, HP-UX has made tremendous progress on its way to more performance, stability, resilience, feature set, and so forth. Thanks to a developer community numbering in the hundreds of thousands, Linux will probably someday get to the same point. We just are not there yet. So at this moment, customers would not dump HP-UX in favor of Linux. As time goes, I would expect that to change.

At the end of the day, it all goes down to very simple product portfolio management rules, that any rookie in any MBA program can tell. The HP e3000 is Phase 4, which is the last stage of decay before an ultimate demise. HP-UX is Phase 3, which is stability, and an enjoyable cash-cow status more or less protecting it from immediate 11/14-like announcements. Linux is Phase 2, which is expansion. I do not yet see a clear Phase One (nascent, emerging, whatever) offering, but it may already exist. It is just a question of life cycle management. Hence, time will tell.

Speaking for yourself, do you see a long-term future for HP-UX, or does it have a November 14 announcement waiting in its future?

Depending on what we call long-term, I may have already answered that question above. In a nutshell, I would forecast a demise of HP-UX when it loses its cash-cow status, which is when it costs HP too much to maintain as measured against the profit it brings in. This may last longer than the HP e3000, due to a now much larger installed base. However, the ultimate demise of HP-UX (which, no doubt, WILL happen) could be hastened by the rise of Linux. Or of yet another unknown technology.

That does not mean customers would not be wise to invest on HP-UX. Perhaps its demise will happen in a decade, or a bit less, or a bit more. Anyway, I would think HP-UX is there to last more than one generation of hardware (3 to 4 years, at current technological pace), thus making it a highly recommendable choice.

What has changed about your daily work since November of 2001? What impact did the HP announcement have on your own duties?

Not much, in fact. HP e3000 business was already declining before 11/14. It simply hastened the slide. What it changed for me is that I have had the opportunity to join OpenMPE, and devote some time to contributing to its activities. I have also been given the assignment to implement a program to retrain our staff, in order to reassign our precious and scarce human resources toward other activities.

So I set up a multi-platform data center, an HP-UX internal training program, and redesigned some of our recruitment tests. Another supervisor has put in place a similar program to refocus resources from IMAGE toward other DBMSs, like Oracle. We are also actively investigating Linux, HP-Eloquence and a few other pieces of software. But, since I’ve created my own business unit to be multi-platform from Day One, 11/14 did not change that much.

One of the MPE strengths is its close integration with the 3000 and IMAGE. Can this kind of advantage be found anyplace else?

The integration between the platform (the HP e3000) and the operating system is going to be de-coupled. The emulator project is just about that objective. So that level of integration is bound to disappear.

IMAGE is different. In fact, one major reason customers one day chose MPE was because the application they had chosen ran on it. And the application ran on MPE because it had IMAGE, which is to this day the most resilient DBMS I can think of. So the integration between MPE and Image can’t be broken. If it were, all the argumentation about a possible future for MPE running above an emulator would become moot, for it would sever the link between the application and IMAGE, which is exactly why people run MPE. So I would gamble for a future for MPE without the 3000, but I would not bet a single penny on a future for MPE without IMAGE.

What else does HP need to announce or create to make the coming Transition easier for its customers?

In fact, the biggest casualty so far after 11/14 has been the trust and faith that users had toward their manufacturer of choice. So anything HP can do to restore that faith is good. Whether is it facilitating access to demo HP-UX servers, facilitating re-training, facilitating financing, is secondary.

One possible gesture of goodwill could be to de-cripple (partially or completely) some software-crippled systems. A few months back, I had the opportunity to run a test suite on an uncrippled A-Class, and the results were outstanding! Now, this is a system! Whether that gesture is useful in technical terms may be next to irrelevant. What’s useful is restoring the faith. I’ve been working with computers for 23 years, and I have never seen a community as closely-knit together as the MPE community. This enhances the necessity to have faith.