NewsWire Q&A: Harry Sterling
Commercial Systems Division
Harry Sterling has a new job to go with the new year. Only a few months
ago Sterling spoke with us in another Q&A, waxing eloquent about the
new contact he'd helped establish between customers and his R&D lab
engineers. With his promotion to Commercial Systems Division general manager
spot on Dec. 22, he gets to promote that Customer First concept, expanding
it to improve the value of the HP 3000 in areas outside of his technical
A New GM for the New Year
That experience is substantial. Sterling headed up the R&D section of
the division for almost a decade, and had worked on MPE/XL migration projects
before that. He came to HP with 12 years of experience as an operator,
and manager of an IBM data center. Sterling was gracious in acknowledging
his technical acumen, pointing out that former GM Rich Sevcik was also a
technically-inclined predecessor. Sevcik held a R&D manager post at
HP's Network Products Division before coming on as CSY GM, and Sterling
noted that division was selling HP 3000 products at the time. But Sterling
may be the GM with the most in-depth knowledge of MPE since it became a
32-bit operating system.
While Olivier Helleboid's departure took some in the division by surprise,
Sterling's move into management follows his growth at HP during this decade.
When we asked for an interview with then-GM Helleboid last summer, Sterling
was also present for the Q&A session. His answers were delivered with
the deliberate expression of an engineer - analyzing and polishing replies
with a direct delivery. Sterling's demeanor on that afternoon suggested
leadership beyond the lab. Now he assumes the helm of a division facing
a challenging year, with increasing competition for enterprise server sales
resulting from HP's growing interest in Windows NT. Unix also remains a
better marketed challenger.
Sterling doesn't see either of those alternatives to the HP 3000 as a threat
to his division's products. Much like Helleboid, he's glad to sell the HP
3000 as a complement to HP's more-marketed Unix and NT solutions. Sterling
is also unabashed about honoring HP 3000 customers' choice of new system
- even if that HP computer isn't a 3000.
Well-regarded by the MPE development community, Sterling is known for his
listening skills. He can use his information gathering to change his mind
dramatically, by some accounts. One creator of MPE products recalled that
"Harry has changed his mind before, sometimes dramatically." The
developer worked with some of the HP team in the customer focused R&D
program and helped set up some visits to customer sites. "I saw Harry
change from his questioning stance into the leading proponent of the
he said. " He does demand hard facts before committing to anything."
While Sterling's "Customer Focus" thinking may have begun with
team play, he seems to approach it from his own engineering experience,
knowing that no system is the right choice for every business need. How
he will make the HP 3000 a better choice during his term, and silence the
doubts about its longevity, were the focus of our talk. We didn't observe
a lot of change in HP 3000 strategy to accompany the change in the GM's
office. We looked for signs that his technical savvy and ease with customers
would get the HP 3000 ready for the next century.
How do you think your experiences as CSY's R&D manager will best
serve you in your new post?
What I've learned with the customer focus, and how important
it is to work closely with customers, will help me most in my new role.
I had contact with customers before, but it was in a very different style.
What we learned in CSY in the past two years was to be in listening mode
rather than tell mode. In the past we had a solution in search of a problem.
That's the major change we've made over the last three years, going into
listening mode much sooner in the process.
How do you balance that listening mode with a need to advocate a certain
solution to the customer?
We're trying to understand the customer and business environments,
then making our investment consistent with that. We look across HP and figure
out what are the ways we can provide the best solution for the
What changes do you want to make under your management?
I don't really see that we're going to make any major changes.
We're going to keep the same structure that we have. We have a good strategy.
We're in the midst of our planning process, and I think we'll keep tuning
the process as we go.
We are no longer platform-centric. We're not necessarily focused on providing
the 3000 as the only part of the solution.
Coming from a division that sells HP platforms, that seems rather unique.
Can you think of any other division selling HP systems that isn't
The others aren't as far along in that as we are, but they're
certainly trying to move in that direction.
How will you assure customers who are worried about the HP 3000's future
now that they're hearing of a 3000-to-9000 HP Conversion Kit?
We've never actually converted a 3000 to a 9000 under one of
these arrangements. I do know that we've converted a 9000 to a 3000. The
customer chose to go back to the 3000 rather continuing with their 9000
This is not something new. We've been doing it for over a year, but it hasn't
been on the price list. In the one case I'm familiar with, they felt like
if they had this guarantee, it would make the purchasing process easier
for them. The issue we're facing is the up-front purchasing decision on
HP 3000s. We're finding some MIS directors get hit with a lot of questions
about standards when they go to consolidate. This kit is simply a way of
making that process easier for them. It became an easier process for the
salespeople if they have this kit available.
Do you think this kit will introduce concern in the customer community?
I hope not. We did this because the customers told us it would
make it easier for them to purchase HP 3000s.
Since you're helping the GSY division sell systems into your customer's
shops with a Conversion Kit, will you push for a similar conversion kit
on the price list to convert HP 9000s to HP 3000s?
If the customers tell us that's what they need, then we'll do
it. Right now that conversion is a factory special, and we can do that.
It's a lot of internal work for us to put it on the Corporate Price
What's your goal for your term as CSY leader?
To further improve our ability to work closely with customers
and to really execute our plans and strategies. Two years from now when
I look back, I'd like to know that we've done the right thing for the customers,
and that the customers are still saying that CSY is the best division to
Why do you think they'll be saying that two years from now?
They'll feel the investments made in products offered by CSY
were a wise investment, and they won't have any regrets in making those
What do you see as the steps you can take to make that happen?
One of the key steps is for me to work very closely with GSY
in terms of integration with our Unix platforms and with NSD in terms of
integration with our future NetServer and NT server platforms. We have to
be realistic about looking at the environments in the customers' sites and
be able to have our products fit very well. That requires that we work very
closely with some of our key business partners.
Describe what you'd like to see happen to CSY's development budget during
your term. Customers are being told there's limits to HP's resources to
work on things like 32-bit ODBC drivers for IMAGE or b-tree indexes.
I think that's true of every R&D organization, don't you?
Aren't there always more good ideas than there are people to work on
Well, there don't seem to be enough resources to meet the demands of
customers asking for those projects.
If you talked to some of GSY's customers and found out what
they want in terms of new features, you'd get a similar answer: we're not
able to do everything as quickly as customers would like.
The bottom line is that every division's customers are always going to
want more than the division can develop?
Nevertheless, there are some fundamental features that have been available
on HP-UX systems that seem to be stuck in the MPE/iX development process:
compilers that can be fine-tuned for different versions of PA-RISC; DCE
capability that's bundled into HP-UX but comes at an extra cost to MPE/iX
customers. Would you like to see a bigger development budget to meet some
of those demands?
Development budgets are set in direct relation to revenue. How
much money our customers spend with us will determine how much we can afford
to invest in R&D. That's the bottom line. One of our corporate objectives
is profitability, and every division is held to a profit objective.
How's CSY's profitability compared to GSY's?
That's not something I can answer. We are both profitable, but
I can't be more specific than that. It's a matter of degree and
How will you change the role of ISO engineers in CSY projects - in what
new areas do you plan to make use of this resource?
I was responsible for setting up that whole relationship when
I was R&D manager, so there are no plans to change anything. The plan
is that they are an extension of our R&D organization and we will use
them wherever they can add value. The one thing I do want to do is to get
them more involved with customers. I'd like to take advantage of the fact
they're on the other side of the ocean and get them more involved with European
Are you satisfied with the core competencies CSY is demonstrating in
its database technology today?
There's certainly more that we could be doing. It's a question
of prioritization and investment levels we're able to make. We are not a
database company or a database division. We are attempting to provide a
total solution for our customers. When we come across a customer need we
will look at all existing products in combination with future investments.
We want to balance all that to come up with the best solution for the customers.
Does the fact that 80 percent of your customer base uses IMAGE have an
impact on that decision to not be a database company? Your predecessor admitted
that databases were a core competency, particularly of CSY's. Do you have
the same opinion?
Absolutely. IMAGE/SQL is a core competency and part of our solution.
And we're going to continue making those investments.
What's your view on the value of b-tree indexing in IMAGE/SQL and TurboImage?
Will you commit development resources to creating a basic-level version
of this feature to include in HP's databases?
We're currently evaluating that, and I don't want to make a
comment on that right now. We'll be making an announcement on that in the
next few months.
Do you think HP 3000 customers are satisfied with HP's database functionality
and co-existence? What are the customers telling you in listening mode?
Yes. The VARs are very pleased with the work we've done with
SQL on IMAGE. I was in Atlanta talking with one of our channel partners,
and they are trying to move their product to use the SQL rather than IMAGE
interfaces so they can be more portable and offer their product on Unix
in the future. They want us to do continued work with performance improvements.
We're going to work with them to better understand how we can tune the SQL
There are other examples where we're working with customers to understand
what we need to do with IMAGE itself - where they're not using SQL. We're
going to keep working with the customers, working through SIGIMAGE, very
actively to understand what the future needs are and continue to make those
Do you believe that MPE/iX can deliver performance equal to HP-UX without
a 64-bit revision?
Yes. The real issue with a 64-bit version of MPE/iX is not so
much performance as it is scalability. We're finding that our larger customers
are becoming concerned about putting all of their applications on a single
box. In some cases they're looking to move to more of a cluster environment,
so they don't have quite the risk of failure of having all of their users
off the system. We're trying to understand if it's better for us to continue
to scale up in terms of the size of the box, or would we be better suited
to work more on a clustering technology - like Shareplex - in making that
a better solution. You could increase performance by adding a new
But on a system-to-system comparison, isn't a version of HP-UX tuned
for 64 bits going to run faster than MPE/iX version that's not tuned for
No, because the clock cycle is still the same. What 64 bits
gives you is larger addressing. As long as we can continue to gain the
improvements of the faster chip, we'll continue to scale performance-wise.
We may not be able to have as many users on the system or be able to run
as many applications as you could with full 64-bit implementation. It's
not clear that's the direction our customers want to go. That may not be
the for the mission-critical environment.
You're hearing customers tell you they want their mission-critical
to be more distributed and start to rely on clusters?
Exactly. We had one in particular that had done a consolidation
onto a single box and now they're backing off to two boxes, because of a
problem in doing backups and in doing system maintenance. Two boxes gives
them more flexibility in shifting their users between them.
This is a change from the traditional MPE/iX advantage of carrying more
users and more applications on a single system than Unix alternatives.
We may not want to go any higher than where our capacities are
already. We may now want to move toward clustering technology. Maybe 2,000
users is enough.
Wouldn't the HP 3000 benefit from having a larger address space? There
are already people who need a greater file size on the system.
That would be the only area that we would actually have an advantage
today by using 64 bits. We actually do have internal 64-bit addressing already
in MPE. We may begin to evolve, and as we need to change limits we may start
to slowly start using 64-bit addressing. It's not going to be a massive,
one-shot project for us. It will be more of an evolutionary transition.
As we need some new capability that 64 bits may give us, we might do
If you started working on it immediately, how soon could you deliver
a 64-bit version of MPE/iX?
We're not going to have a massive 64-bit conversion. We'll evolve
as we need to take advantage of that. Right now I'm not sure we know where
we need to use 64-bit and where we wouldn't.
Are we still looking at a major release with significant new functionality
every year or so?
About every nine to 12 months. We're still looking at platform
releases every 18 to 24 months. We're already planning the release after
What's your goal for new installations of HP 3000s?
I don't have a specific goal other than to continue working
with our strategic VARs and VABs, which is primarily how we sell our systems.
What do you see as the essential, can't-be-duplicated advantages of owning
an HP 3000?
One clear differentiator is ease of use. It's still much easier
to use a 3000 than a Unix solution.
I'd like to think it's the relationship that customers have with CSY more
so than the platform. Our objective is to have the customers believe when
they spend their dollar on an HP 3000 they're getting the most for that
dollar. It's a matter of protection investment and lasting value. Our customers
don't see that when they purchase another platform. If our GSY friends get
to that level, then I think our customers win overall.
What changes do you want to make to communicate these advantages to new
customers and those who are defending their 3000s to upper management?
We're always looking for ways to get the message out to customers,
to have a two-way dialog. I'll continue to keep the organization focused
on customer contact and communication.
What areas of business computing do you want to see the HP 3000 become
stronger in during your term?
Co-existence, to have the 3000 exist in whatever evolving
the customers have in their business climate. We haven't done a lot yet
to understand from the R&D side what coexistence means from an NT
Will coexistence bring the 3000 into shops where it's not installed
It could, by working with some our VARs who are positioning
their solutions to a new customer. If their customers know the new 3000
solution could easily link into their existing environment, it could be
a real advantage.
Can you make a case for a single-user, development HP 3000 priced for
small companies to use in building MPE/iX applications? This would be a
system bundled with a development environment much like SoftBench, priced
in the same range as an HP-UX workstation.
No. I'd rather work with those companies individually, to try
to come up with a kind of deal that would let them have that kind of capability
at the lowest cost possible. If we tried to use a Unix workstation as a
platform, there's no NIO bus for us to have I/O with, so we'd have to write
all new I/O drivers for the peripherals that would be integrated in the
box. We'd have to leverage an existing box. It would be a huge investment
for a very small payback. I don't think the cost of the development platform
is an issue.
What steps do you want to take to increase the number of HP 3000 host
applications during your term?
We're attempting to make the 3000 as flexible as possible, to
meet the needs from both Unix and MPE directions. Our coexistence strategy
does provide the customer with the applications they need, without having
to change their current environment. The installed base is growing today,
because we have very strong third party VABs selling new applications to
We're not being real successful in terms of attracting new applications,
and it doesn't have anything to do with the platform itself. It has to do
with the VARs ability to support multiple platforms. That's why we've moved
to the coexistence strategy. We're done our work in the past couple of years
in trying to attract many new applications to the 3000. The bottom line
issue for a VAR is what's the largest market opportunity for them to sell
their solution, and how much does it cost for them to support their product
on multiple platforms.
Even though we have made the 3000 compatible in terms of portability, it's
not a platform that most large application vendors are willing to support.
They're even limiting the number of Unix platforms they're
What's the most damaging myth about the HP 3000 you want to dispel during
your term as General Manager?
That the 3000 is going away. That's the myth that I want to
go away. I want customers to believe and know that when they buy a 3000
it's a good, wise decision when it's best suited to meet their needs. We
in CSY and HP are going to be here to support them well into the future,
to ensure their purchase is of lasting value and that they can evolve to
the future, whatever that is.
Mail your comments to Sterling on his answers
back to the NewsWire Planning table of contents