Tending a Steady Light for Support

Mike Hornsby has a longer view of the 3000’s future than HP does — partly because his company has a clear look at the computer’s past. Hornsby heads up Beechglen, a company dedicated to support HP 3000 customers regardless of the vintage of their computers. Beechglen doesn’t mind who calls for support, or what version of MPE the customer’s got in place. Its “answer any question from anyone, anytime” approach even extends to software and hardware outside of the HP 3000 realm, a remarkable open door approach to support that Hornsby says his clients have respected, despite the potential for abuse.

It’s also an approach that springs from Hornsby’s roots as a Hewlett-Packard systems engineer. He served HP 3000 customers in an era when in-person support was commonplace, and HP SEs called to see how you were doing, not just to resolve problems. He specialized in HP 3000 networking and performance as a Senior Systems Engineer in HP’s Cincinnati, Ohio office, then founded Beechglen with the wife Mary Jo in 1988. He said he left HP because the customer contact he enjoyed most was being transferred to HP’s telephone Response Center. More than 13 years later Beechglen has a group of HP experts who team-solve problems when needed, sporting a reputation for being able to answer almost any 3000 question. The support keeps enterprise-level sites running on software and hardware which HP has long ago given up on.

It’s that last feat which brings Hornsby to our interview space this month, more than 60 days after HP’s announcement that it’s leaving the 3000 business by the end of 2006. Customers who don’t march in lockstep with HP’s recommendations might be wondering who can serve their support needs beyond that date, or even now at a better value. Beechglen is just one of many companies making that pledge, but it’s one of the largest with a deep 3000 background. We asked Hornsby how much of the 3000 customer base Beechglen’s experts might be able to cover now that HP is conceding the territory, and how he views the prospects for continuing to use the 3000 and get it supported, too.

How did you get the idea to provide an alternative to HP’s support for 3000 customers?

Beechglen evolved into support services after we partnered with Kelly Computer Systems. At the time they were the leading sellers of non-HP add-on memory for HP 3000 systems. We would provide performance tuning and analysis to those sites where the additional memory alone did not provide sufficient performance gains. We became the path of least resistance for support questions for these customers, and we worked out an ongoing relationship enhancing or replacing the phone in support from HP. The support philosophy was always to “take care of the customer” by providing a “phone a friend” for any software related issues.

How many are you helping now?

We provide assistance for the entire HP 3000 community in a number of ways. We have a direct telephone support service, but also provide free assistance to anyone via the 3000-L mailing list and newsgroup, other miscellaneous newsgroups, the Interex AnswerLine, and even at HP’s ITRC forum on the Web.

Are you willing to ramp up to support a lot more customers?

We have been in an active growth state for the past several years. Y2K conversions and testing created many support opportunities, and since then, many more HP 3000 and HP 9000 opportunities have developed.

What do you consider to be the special experience that Beechglen brings to this kind of work?

Our goal is to take care of the customer, as if we were them, knowing what we know. Our motto is ‘anyone can call about anything anytime. From the beginning we have realized that it is much easier to answer a question at 2:00 AM than to cleanup after a guess at 8:00 AM. Our support is immediate, like having a ‘phone a friend’ who can answer the question or solve the problem more than 95percent of the time on the first call.

For example: A customer calls asking how to stop his laser printer from switching paper trays when the first tray is empty, because the second tray is letterhead. Our consultative answer was, here is the escape sequence to label the trays so that the printer knows what is in each tray. But, it is also possible to print letterhead from the HP 3000 at no additional cost using environment files. The elimination of preprinted forms saved them the printing cost and the operational overhead of stocking and switching the forms.

Another example was a customer who was frustrated at having to dial in every night to check in on a credit card upload job. We showed her how to use a freeware e-mail client from Telemon to e-mail her pager if the job failed. The special experience is that we take care of our customers as if they were family.

People get concerned about the availability of hardware in the years to come. Has it been your experience that HP’s parts supply is essential to repairing systems?

Certainly parts availability and staging is key to providing 4-hour onsite repair. Fortunately there are many excellent third-party service providers available. In our experience they are just as good and in many cases better than HP’s service. In just about all cases this is their main business, and they are totally focused on parts and service. Many of our mission critical ‘always on’ customers also have hot backup systems. With HP’s 11/14 announcement this will become more feasible for many more HP 3000 sites.

What about patches? HP probably won’t be writing many within a few years. How could you provide support for a customer who needed some part of MPE patched to solve a problem, if HP’s not patching them anymore?

First let me say that I have always been skeptical about patches. To paraphrase a famous coach, “three things can happen when you patch, and two of them are bad.” Our approach has always been to be more consultative and look at the problem from a programming prospective. Most of the time people run into bugs because they are off the beaten path, and trying to do something the hard way. We steer them back and find a better alternative. This approach minimizes down time and risk of introducing newer more serious problems.

We have also created some of our own patches. We had several customers who required an extended security log in prompt, so we created LPRESET. Some good information on these types of things can be found in an article I wrote for Interact, “MPE Debug and Dump Analysis.”

We keep hearing stories about how Beechglen is more responsive than HP’s support. What kinds of service response levels do you offer, and can you compare them to HP’s?

The only type of support service we provide is 24x7x365, so that “anyone can call about anything anytime.” Our pricing is still significantly less expensive than HP’s 8-5 coverage.

HP’s phone in support is limited to very specific, fairly closed-ended questions, related only to HP products, and was conceived and designed at a time when the average system manager was more of a specialist on the HP 3000. Today, the average system manager is more of a ‘Jack or Jill of all trades,’ part time system administrator, part time network administrator, and part time PC help center. They have many items in their ‘to do today’ list and most of them are numbered 1.

We help to offload some of those tasks, take the ball and run with it, reporting back on the results. This is especially effective for the kinds of tasks that are only done once in a great while, like adding a network printer or disc array. It might take a system administrator several hours to complete this task, whereas we can access the system remotely, and accomplish the same task in minutes.

Is there really any further reason for a 9x7 owner to be paying for HP’s software support on their HP 3000 systems, now that 6.0 is being removed from HP’s support in October?

No, for several reasons that may also apply to other HP 3000 models. First, software support is so expensive, so that for many models you could buy an entire matching system with a current license for less than one year of software support service.

Second, at this point it would be a mistake to take a system that is working and update it for the sake of HP’s version support timetable. I would strongly recommend a strategy of postponing updates until a specific business need requires it. Finally, 6.5 and 7.0 have features which extend the high-end of the HP 3000 at the cost of degraded performance for the low-end systems.

HP wants to portray the 3000 customer base choosing to Homestead as less strategic about the way they use their 3000s. Is that your experience among your customer base — that the 3000s aren’t the most important in those IT shops?

If I understand your question as: HP wants the HP 3000 base to move to other HP platforms ASAP, and those that don’t share this view are somehow strategically defective? Then I have to disagree with that. We support both HP 3000 and HP 9000 systems and I personally made over 10,000 calls to these system administrators in the past year. I would estimate that less than 5 percent of the remaining HP 3000 installed base would even consider another HP platform.

Another way of saying this is: If you had a 10-year-old car that was paid for and was extremely reliable, would you trade it in just because that model was no longer going to be sold new? No, because the new car can’t be more reliable — and for the expense you won’t get any marginal benefits.

Reading strategically between the lines of HP’s 11/14 announcement, one item seems very significant. Hewlett-Packard has basically admitted that they could not justify the effort to convert MPE/iX, IMAGE/SQL, and various subsystems to run on IA-64. This does not bode well for other complex operating systems, DBMS, and application packages, attempting the same “migration.”

My recommendation is to ride out the wave of the coming 64-bit storm, and see where the chips eventually fall. Meanwhile, your current HP 3000 investment will continue to pay reliable dividends. Otherwise, customers could be faced with an expensive initial parallel investment, immediately followed by an equally expensive 64-bit upgrade. Another way of putting it would be that it would be a shame to jump out of the frying pan into the fire.

Would you say that HP’s continued support for IMAGE is more essential than its 3000 support in general? Or would you recommend that customers start looking at moving their IMAGE data into other databases? How can third-parties step into support for the 3000’s database?

The secret to the success of the HP 3000 has been that IMAGE and MPE were integrated together. This provided the ability to obtain the reliability and performance standards the HP 3000 attained. So in my view the two are inseparable.

I think you’ll see the remaining installed base split into six categories:

1) Those with application packages that have financial resources to switch to a different package: These will switch gradually through 2006 and the switch will take a relatively short time.

2) Those that have complex custom code that have financial resources to switch to an application package: These will switch to application packages through 2006 and the switch may take 3-5 years. Many HP 3000 sites are already in this phase.

3) Those that have simple custom code and budget to convert to different platform: These will wait as long as possible for a stable alternative.

4) Those that have no budget: These will also wait as long as possible to choose one of the three alternatives above.

5) Those that have extremely complex applications that were evolved over 20 years in many different generations of programming languages: These will take years just to document what the existing system does, and at least another 10 years to accomplish a complete migration.

6) Those already in archival mode. The HP 3000 is for lookup and reference purposes due to statute or policy regulations. Overlooked is the fact that many HP 3000 applications will by law be running well past the end of 2006.

If you could only get one more significant IMAGE enhancement out the CSY labs, what would you hope they would do?

I would like to see many improvements to IMAGE; unfortunately there is not enough time till December of 2003 for any enhancements to make it into a release, then to be installed on enough systems, to get enough patches to make the enhancements recommendable for a production system. Keep in mind that 12/31/2003 is now closer in the future than 12/31/2000 was in the past.

What’s your outlook on the prospects of a company running home-grown applications on a 3000 to keep their apps on MPE, given the Nov. 14 announcements? Is there enough help out there in the 3000 community to let them stay successful with the 3000?

Definitely. I think you’ll see a major percentage of the current HP 3000 base still in production usage well after 2012. We still have several customers running MPE/V on a wide variety of those models, some that have been off of HP support for more than 10 years.