Late last week,
I had the opportunity to spend some time on the phone with a customer who had a
large UNIX install base, but was considering HP and HP-UX 11i v3 for the first
time. The account team was on site, and obviously covered many more topics than
I did during my 45 minutes on the phone. However, I found the customer's
response to what I had to say quite interesting, and it got me thinking about
server design, and real-world customer problems.
While I covered
the whole software portfolio, the customer really had questions about two areas
- our high availability and disaster recovery portfolio, and our orchestration
capabilities. In fact, I ended up going back to our orchestration products multiple
times, just to answer all of their questions. The customer was amazed that HP
Insight Orchestration, as part of our Converged Infrastructure, could automate
such a big part of their day to day provisioning process. To them, this would
help solve a huge operational problem - reducing the ongoing maintenance costs
as part of their overall it budget. This is something HP experienced - reducing
maintenance and operations from 70% of the IT budget and reducing it to around
30% of the IT budget. The first model was making IT unsustainable.
This got me
thinking about the what our customers ask for in new servers. Everyone expects
that the new servers will offer more performance than the previous generation.
Everyone expects that power consumption will drop. These are good
characteristics, but is that all the innovation you want from your server
vendor? A faster chip, but in the same old box - what comes to my mind when I see the new POWER 7 servers.? With the same old maintenance
issues? A vendor who makes great hardware, but looks at customers as support
annunities? A vendor who isn't invested in reducing maintenance as percentage
of your IT budget - solving the big IT problem?
released their new Itanium 9300 processor, and HP will announce its new systems
within 90 days. My question: will they just be the same old servers, just with
a faster processor? Or will they be something more for mission critical
environments? Will HP take steps to help our customers dig out from their
maintenance costs? What are you looking for in these new servers?
I caught parts of the Oracle + Sun web cast last week. Like many people, I've been thinking about the content of the web cast. In particular, Larry Ellison and other speakers mentioned moving back to the concept of integrated systems, much like in the mainframe era of the 1960's.
While I don't pretend to know much about the mainframe market in the 1960's, one thing that I do know is that the way IBM ran their business back then lead to alternatives such as the mini-computer, and ultimately paved the way for UNIX systems, still often referred to as Open Systems. The market is much different today, with multiple strong vendors, including HP, offering customers choice and driving healthy competition. Let's face it, competition keeps everyone on their toes!
Apparently, I'm not the only one questioning how great the mainframe market of the 1960's really was.
John Rymer at Forrester has a great analysis on his Forrester Blog at http://blogs.forrester.com/appdev/2010/01/peace-love-and-the-ibm-system-360.html that covers this topic rather well. He has an interesting list of the pros and cons of the mainframe era that makes for good reading.
In addition, I know that in my discussions with customers, Oracle and Oracle licensing is a very popular topic. Integrated solutions definitely help drive down the maintenance cost for that solution. However, standardization across multiple projects also drives down costs. While there are always good business cases for one off exceptions, I know that keeping those exceptions to a minimum helps drastically reduce backend support costs. It will be interesting to see how this plays out over the next year.
This is one of the first changes that will impact the UNIX business this year, and I'm sure that there are plenty more coming.
Well, it is a big day in the UNIX industry - the Sun acquisition by Oracle has been completed, and the UNIX landscape is changing.
I caught a bit of the Oracle + Sun webcast between calls. I found that it was really interesting that the first part of the web cast was about reducing the risk on change management, integrating solutions, and at the end of the day, reducing the maintenance costs. I'm sure that some of the Oracle + Sun solutions will help do that.
Having said that, this is something that HP has recognized and been working on for years. The HP Converged Infrastructure message starts with the idea that we want to switch from 70% maintenance and 30% innovation to 30% maintenance and 70% innovation. We also have a great proof point - at HP, we did this with our own internal IT, a project that finished a few years ago. While HP and Oracle agree at a high level, the implementation may be different. It will be interesting to see how this plays out over the next day, weeks, and years.
One of the reasons
that I came home between the training events in London and Frankfurt was to
cover a customer briefing at the Cupertino Executive Briefing Center -although
it also saved HP money, and let's face it, everyone's budget is still pretty
The briefing was an
overview of the HP and HP IT Transformation. However, there was a specific
deeper dive and conversation around our desktop management - everything from
asset management, self serve processes, to help desk processes. While I often
speak about mission critical environments, it is great to also have a glimpse
into the desktop and notebook environments - especially since as a user, I'm a
customer of our internal services.
It's been a two
years since I last got a new notebook, and I got to use some of the new
processes in person. Instead of waiting for my business to have available funds
for new notebooks, we now get automated e-mails when it is time to refresh our
systems. We select our user profile, and a week or two later, the system shows
up at our desk.
Then, instead of
someone from HP IT disappearing with our old and new system for half a day or
more, we log into the new notebook, finish some minor set up ourselves. Both
systems are then unplugged from the network, software is run on both systems,
and all of our settings and data are copied to the new system. Even better, I
can do this at home at night, so I'm not sitting around the office looking
useless, waiting for my systems to come back. I have much less downtime, and IT
doesn't have someone spending a lot of time just copying data from one system
to another. Even better, they have made this process easy enough that most of
my co-workers who normally call the helpdesk (or me) whenever they run into any
sort of problem now do this on their own, with no help.
But the interesting
part of this past year is seeing the impact of the economy on our help desk.
Due to the poor economy in 2009, HP pushed out the refresh on some systems by
another 6 months. It turns out, that they can see a direct co-relation to the number
of help desk calls as they did this, something that we could never track in the
past. Users of systems that are older than our usual refresh age are now
logging more calls, more downtime, and more repairs than they were in the past.
We helped save the company some money during the economic downturn, but it is
definitely driving up our maintenance budget. Since our goal here at HP is to
keep our maintenance costs low, you can bet that this will change as the
economy becomes more predictable.
However, the biggest
difference today compared to a number of years ago is that we can now track the
impact of decisions like pushing out the refresh rate for desktops and
notebooks, something we never could do in the past.
This begs the
question about what happens in a mission critical environment. There are lots
and lots of mission critical systems out there that have been in place for 5
years, 6 years, 10 years, or more. While the server life cycle definitely is
longer than desktops and notebooks, at some point, we run into the same issue -
old systems cost more to maintain than new systems (especially after the
warranty expires). Just because the system is still up and running, and has
been paid for, doesn't mean that it is cheaper than buying a new systems. We're
figuring that out at HP for our internal use.
Does this apply
outside of HP? Can you actually track the relationship between system age and
maintenance costs? What have you found?