| Front Page | News Headlines | Technical Headlines | Planning Features | Advanced Search |
 Adager graphic Adager Sponsor Message

Terry Simpkins
Director ISIT
TRW Automotive Electronics
Sensors & Components

 

It’s All About the Uptime

Terry Simpkins knows how to focus while meeting his computing challenges. One of the founders of the ASKUS manufacturing user group for MANMAN, Simpkins has managed HP 3000 shops for more than 15 years — his current post is that of IT director for the Sensors & Components Division of TRW. His slice of the TRW empire is a $200 million enterprise flung across 17 time zones and 10 plants in the US, China and various parts of Europe. All of the Sensors & Components operations are run off a single HP 3000, a resource shared so judiciously most users don’t know they’re computing alongside a handful of other sites.

What makes Simpkins noteworthy is his success at keeping the HP 3000 at the hub of such a widespread operation, moving manufacturing from acquisitions onto a platform that’s proven and reliable. This accomplishment isn’t easy in a world where vendors claim platforms don’t matter as much as applications — a place where churn is the order of the day from the mainstream IT media. He’s most recently testified for the platform in the first cost-of-ownership brochure produced by the 3000 division, saying the reason his company has stuck with the system is “it just works.” He’s doing that while employing Windows NT as a mail system server with file and print sharing duties, so it’s not like his testimony is without comparison.

Simpkins began his career with the HP 3000 in 1981 as a programmer-analyst at HP’s Disk Memory Division, maintaining an inventory control system. In the 1980s he grew a $10 million manufacturer’s IT operation to handle business of $100 million at Spectraphysics, then left to consult while looking for his next post. His current position has him directing an IT staff of 22, and he’s been through a corporate acquisition when the TRW group purchased Lucas Verity.

We think Simpkins stands as a prototype of the modern IT director who’s putting the e3000 and MPE in the driver’s seat for his corporation — someone who’s defended the business sense of the 3000 through corporate change, politics and technological fashion. In an era when the 3000 is getting little mention outside the system’s installed base, we wanted Simpkins to tell us how modern IT manager can keep a company focused and migrate other systems onto its 3000. His answer seems to be, “Stay focused on the platform that delivers the best uptime.”

You’re known as an IT director who’s moved manufacturing systems from other platforms to the 3000 and MANMAN. What platforms have you converted from to a 3000? How do you sell that concept to your management?

All of our manufacturing plants were run as autonomous businesses. We had six MRP systems, six order entry systems and five general ledgers, running on five different platforms. We had PDP 11s, Prime, an IBM System 36, an IBM mainframe, one PC-based application. The HP 3000 applications were GrowthPower and MANMAN. As we built this business from a group of autonomous businesses into more of a cohesive business unit, one of the things we had to do was to have common business systems and architectures.

We decided to run on the HP 3000 and MANMAN. As we went through the process we would migrate those people. We sold it on capabilities, price and performance. We were able to show them that while MANMAN on a scale of 1 to 10 wasn’t a 10, it was certainly a 7 or an 8. We were able to convert eight manufacturing facilities over to MANMAN in about two and half years for under $2 million. That included hardware, software, licenses, training, consulting and travel. If you were to look at the new, snazzy, sexy SAP or Baan, or Oracle, you started at about $4 million and went up from there.

We started the process about four and half years ago, and there was a critical shortage of that kind of Baan-SAP-Oracle talent in the world then. The standard horror story is that you had consultants in and you’d change them out every six months because they’d get a better job offer, take it and leave you high and dry. We said we didn’t see that as a plus. The one of our sites that was running MANMAN was running it well and was happy with it. We asked them about it and they said, “It just works. It runs our business.”

There was nothing overly complicated about our manufacturing processes. We adopted a strategy that said we’d go with a known, proven, reliable product that we believe we can get sufficient resources to run, both on the people and the hardware side. We had experience on the 3000 and knew how it worked. We had experience on MANMAN and knew how it worked. There was a fair amount of believing in “the devil you know versus the devil you don’t know.” We felt we knew where most of MANMAN’s shortcomings were, so we wouldn’t get blindsided by them.

My management didn’t have any attraction to having the latest and greatest. They said they wanted to devote our capital and resources to making our business better, as opposed to having better systems.

HP likes to tell us in the press that IT managers at your level don’t make deployment decisions around platforms anymore, that applications are the only thing that matters. What do you believe?

I think that’s bullshit. If I’m looking for an application and I find two that run equally well, and one of them runs on a platform I already have expertise with, I’m all over that one. I don’t believe that we’re all in a heterogenous environment, or that we want to be in one. I’m not afraid of a heterogenous environment, but why do I want to add complexity to my life if I don’t have to? I think HP’s selling that because it’s a way for them to tear down a barrier to entry in businesses they’re not in already. If somebody says we can do this equally well on a 3000 or an NT server, my initial reaction is “Why wouldn’t I want to put it on the 3000?” I may come back and see that it may put me over capacity on the 3000. But then I’ll look to see if it’s cheaper to up my 3000 capacity or to put an NT server in.

That last question could easily tell me it’s cheaper to put an NT server in, because I won’t have to pay the license upgrade fees on the 3000. But if I can put a separate 3000 in and run it from there, as opposed to putting in an NT server, you dilute that argument, because the cost is the same. With some of the used 3000 boxes of today, the costs are about the same. The prices have come down dramatically on them. I’ve already got the resource, I know how to use it and use the tools on it, I can buy second copies of tools I need and not have to pay first-copy prices. Again, it comes down to a devil I know versus a devil I don’t.

How are you handling the tight labor market for IT people in general, and are you experiencing a shortage of experienced MPE staff? How are you training for the platform specifics?

The market is a little bit tighter than it was four or five years ago. But I don’t perceive it as a reason to change platforms. I haven’t had too much trouble hiring IT people. We feel a lot more comfortable hiring MPE people, because we feel like we’re able to judge well and get good ones. I’ve had several instances where we hired PC and NT people, and we got a little buffaloed. We’re not as good at weeding out the bullshit from the truth with them. Since our background wasn’t as strong in that area, we weren’t able to differentiate the truth. I’m focused on getting a little bigger, so we can afford to bring people in and train them on the 3000.

We haven’t been able to do a lot of training. Ours lately has been limited to getting people to user group meetings and sitting in on MANMAN training when we were doing it for users.

Is there 3000-specific training you want could find and use?

I don’t feel too constrained by the availability. I feel more constrained being able to bring in people who need the training. The problem has been that we’re understaffed and under the gun, and haven’t had time to do some of the training we’d like to do. We want to bring in some sharp people who just haven’t been on the 3000, and bring them up to speed. I’d set them with somebody on my staff. I probably wouldn’t send them to the HP system manager’s class. I’d be more inclined to send them to a Cognos class for a week. If they’re a programmer who’s just unfamiliar with the platform, they can get what they need. If I was hiring somebody in from college — and I don’t know when the last time was I did that — I’d consider sending them off to HP’s intro programming class. They’d go through a real good learning process on the programmer’s intro.

What’s the advantage to your company in sticking with MPE for mission-critical apps?

We looked at it from a price and performance curve, and said we were running 350 users each day on a 959/200. That system is pretty cheap these days. I don’t know of anybody running that many people on Unix box that’s the equivalent of a 959/200. When I was consulting we were directed to write some stuff on Unix, and we found that with five of us doing development, it took one full-time person to be the system and database administrator. I don’t have anybody dedicated to that. I have a system manager, but he works on utilities, menus, takes care of the 3000 operating system upgrades and some applications work and handles payroll. My gut tells me that he’s maybe half-time on the system, and spends virtually no time on the IMAGE database. You just don’t do anything to IMAGE, so you don’t worry about it.

We know the platform, so we stay with it. It’s doing a hell of a job for us. I don’t remember when the last system failure was.

Is there any downside to you or your company in staying committed to a platform not on the lips of many analysts or industry peers?

It never has been on their lips. The 3000 has never been on the radar screen for the industry. I think there’s a lot of hysteria or paranoia about that. I share it, and it frustrates me when you look at how loyal the customer base is and what great stuff the 3000 has, and HP just tends to ignore it. I think it becomes a bit of a self-fulfilling prophecy that HP tends to talk about the things they think the media will quote, and the media is always a driver of thrash and change. It’s what they do by definition, except where you have media outlets like the NewsWire that have a focus. The general media — InformationWeek, PC Week, pick the rag of your choice — thrive on thrash, because if they’re not churning the water, what are they going to write about? They have a vested interest in controversy and stirring up change.

It pisses me off because HP won’t talk about a very large group of very satisfied users. You’d think they’d throw them a bone occasionally and talk about the 3000. I’m inclined to believe they’re doing it intentionally, because it’s been brought to their attention so many times. I find it very telling and borderline offensive that [CEO] Carly [Fiorina] isn’t making a personal appearance at the HP World show. It might be that I’m an ignorant hick and she’s a busy person. But as I like to tell people, I’m the customer and that’s my perception, therefore it’s reality.

I don’t think [staying on the 3000] it’s a big risk to my career. It may limit my career as far as the breadth of opportunities available to me. But I’ve made that choice. And in all honesty, the 3000, on a day-to-day basis, is less relevant to my career than the fact that it’s up all the time. That’s much more relevant. The fact that it’s a 3000 is irrelevant, just so it’s up. And it’s up.

And your operation serves 350 users and 10 plants with just one HP 3000?

Yup. To be fair, it really helps that we’re scattered over 17 time zones, so I never hit it with everybody all at once.

How much in-house development staff do you need to keep your 3000 resources running? Are your sites doing much coding for the 3000?

We have a strategy that says you run MANMAN as it is. We run it vanilla, with nine modules and 15 mods. They’re all reporting-type modifications. I don’t believe we’ve changed the transaction flow of MANMAN anywhere.

We have said it’s better to modify our business processes to fit the way MANMAN works. It’s pretty doggone flexible, not like Baan, which says, “You will do it this way.” They tell you what your processes will look like. MANMAN doesn’t, and it provides a substantial amount of flexibility. We’ve said we’ll modify our processes to the way the system works so we don’t spend lots of money on system development. The old adage still holds pretty true: For every dollar you spend on development you’ll spend $10 to $15 maintaining it over the life of the code. A new system that only costs $100,000 means you just committed to a million dollars.

We have about four people who can touch MANMAN to modify it, and we don’t do it. We will clone code occasionally if there’s a report someone needs and it needs a few fields added to it. We want to be able to stay reasonably current to migrate to new versions.

MANMAN is no spring chicken, in terms of applications. What’s been key to your success in keeping it capable of handling your company’s business?

We recognized the warts that MANMAN has. It doesn’t come standard with a graphical interface. That doesn’t really do a lot for you as far as productivity. I don’t buy the argument that it makes you substantially more productive. If you put a couple of bolt-on [software products] onto MANMAN for not a lot of money you can get partial key lookup capabilities. You can dramatically improve its reporting capabilities with a couple of tools.

Who does the integration when you use bolt-on software for MANMAN?

We rely on whoever we’re buying from to provide something that does bolt on. And we spend our time making sure it integrates correctly and works. If it doesn’t we don’t modify the [MANMAN] code, we go back and tell them we found a bug and don’t think it’s integrating correctly.

Have you found the need to make much use of the newer MPE features, like FTP, Java, Samba, or the Apache Web server?

We’re trying. We’re finding a growing need to use ODBC to extract data out of MANMAN and put it into spreadsheets, for some bolt-ons our users have created. They’ve created some RMA systems and want to get the data right out of MANMAN. We go get it with Minisoft’s ODBC.

We use Minisoft for our terminal emulator, and got a worldwide license when nobody else was offering one. We also use their NetPrint product. I typically try to have fewer vendors and use more of their products. We try to leverage our relationships. On a selfish note, it tends to make me more important to them, because I have more of their products.

One of my guys is off looking at Java, reading up on it. We’re trying to find places where it might be a good tool for us. I’m very glad to see that it is ported to the 3000. It goes a long way to helping me maintain more of a homogenous environment, because I can use those tools on the 3000. If I find something Java can do, I can run it on the 3000. At least I have the option; I don’t have to go get NT.

What level of public commitment do you want to see from HP for the 3000? Is what the CEO or other HP corporate officers say about MPE important to you or your career?

I don’t think it’s got an impact on my career. But it depends on your boss. My boss trusts me. He hires me to know and asks me questions, and doesn’t second-guess me. I’d have a very hard time working for somebody like that.
I’d like to see HP use MPE in the same breath as they say Unix, Linux and NT. People don’t know every since Hertz rental car transaction runs on a 3000. Why don’t the beat that drum, or about OpenSkies? They argue that people don’t care [about platforms]. Bullshit. People care. Those customers put those things on the 3000 for a reason. I suspect that it’s because the system is scalable, reliable and secure. So why not tell people that? I think they’re doing themselves, and me as a shareholder, a great disservice.

 


Copyright The 3000 NewsWire. All rights reserved.