I recently polled several colleagues and Grid evangelists from both the
vendor and end-user communities to gather a fresh perspective on the
continued state of confusion surrounding the rate of Grid adoption and
just how to size this market. The responses I got were fairly
consistent and -- no surprise to my readers, I'm sure -- notably laced
"Trying to discuss how Grid fits into the enterprise can be as
frustrating as squeezing the proverbial square peg into a round hole."
"The problem in trying to understand the potential market for Grid
deployments is that we're trying to compare apples to oranges ... and
bananas ... and grapefruits ... and lemons."
"There are way too many definitions for Grid -- we need better standards ... and industry-standard definitions."
And, my favorite:
- "What a crock ... that's not even a Grid!"
According to Robert Fogel, director of worldwide Grid strategy and
business development at Intel, it is fruitless to debate what is a Grid
and what is not a Grid.
Fogel: "A far more beneficial debate would be a discussion of which use
cases should be considered within scope and which ones are out of scope
for a given specification or product."
Fogel believes that blaming the state of confusion surrounding the Grid
computing market on a lack of standards is "putting the solution cart
before the problem horse."
This writer couldn't agree more. But, I wanted something more concrete.
I wasn't looking for the cart, I was looking for the horse. And I
wanted to hear what was going on -- straight from the horse's mouth.
So, I made a few calls to see who might be interested in sitting down
with me for a bit of a secret chat -- a candid discussion with no hype.
It's not hard to find the big-name end-users of Grid computing these
days. The vendors quite proudly put them on display, and then work hard
to keep them loyal. It's a beautiful thing.
And, like most of us who enjoy the luxury of being backseat drivers, I
have some pretty strong feelings already about the winners and the
"wannabes" in the Grid market. But, thinking back to some pretty
exciting years of HPC innovation, I was wrong once before (ok, maybe
twice) -- and I still own a wheel barrel full of that worthless stock.
So, I must admit, I tend to be a bit less emotional and a lot more
skeptical these days.
I know how the marketing drill works. I've been on the other side of
the fence, working with customers who know it is in their best interest
to paint a rosy picture while they hide their Band-Aids and scars from
the public. If things in the Grid market were really as good as some
would like us to believe, we'd all be relaxing at some Caribbean resort
sipping cool drinks. The square pegs just don't seem to be fitting well
in those round holes. I was dialing for answers.
After just a few calls, I had a taker. And while I promised to keep
this person's identity anonymous, I don't think the lack of an actual
name will detract at all from the impact of this discussion.
So, there we were ...
It was a Sunday morning. 10:30 a.m. Just a little chilly with a
lingering freshness of the light rain that had fallen the night before.
I arrived about 10 minutes early. Our rendezvous point was a tiny
French coffee shop located in an almost hidden alley in a city that we
won't name. I went in and ordered a large house coffee, with two shots
of espresso, and room for lots of cream. With my steaming mug in hand,
I strolled outside and grabbed a table that seemed to be offset just
enough to give us a sense of privacy. Feeling the chill in the air, I
walked across the street to my car to grab my windbreaker out of the
trunk. When I returned to the table and sat down, I realized I hadn't
cleaned out my jacket pockets since last wearing it for a round of golf.
I pulled out several golf tees, a $20 bill, a cigar cutter, a lighter
and two of my favorite cigars -- La Aurora Sapphire Preferidos -- in
their haunting blue canisters. I stared at the cigars. How could I have
forgotten two extremely rare cigars like this -- and not returned them
to my humidor?
As I was pondering my own state of absent mindedness, my guest arrived.
We'll just call him Mr. Preferido. We shook hands and he slapped me on
the back and said, "Alright -- cigars and espresso to start the day. I
He strolled inside to get some coffee and I thought, "Sure. Why not?" I
had just recently been reacquainted with the enjoyment of a good cigar
and a strong cup of coffee. And, from experience, I know that many
executives seem to be quite reflective when enjoying a good cigar.
Ten minutes later, we were puffing on cigars, talking about golf and
slowly working our way around to the topic of Grid computing. But,
eventually, we got there.
I should point out that my guest wields a pretty heavy pen. His IT
budget is a mid-cap company all in itself. I turned on my tape
recorder. We talked. We sipped. We puffed. It was a great day.
Mr. Preferido: "Mike, the problem I see with a number of my peers is
that they have been handed this powerful new tool (Grid), and no manual
on how to use it. So, they wrap all their thinking around doing the
same old job they always did -- just trying to use a new tool -- when,
in fact, they should be thinking of new ways to do the job now that
this tool called Grid computing has taken away many of the limitations
they previously faced."
He puffed on his Sapphire stogie, watched the smoke slowly rising to the top of our bistro table umbrella and then he continued.
"Now, I find this fascinating -- and relevant. Six out of 10 senior
executives surveyed at a recent focus group said they now use some type
of electronic, hand-held personal assistant to help manage their lives.
Several indicated they were completely organized and running their
lives from their cell phone devices. Is it interesting that four out of
10 executives are still holding out and relying on hard copy day
planners and some level of support from a trusted personal (human)
assistant? Or, is it more interesting that the 60 percent indicated the
use of new (PDA) technology for planning and organizing their lives has
helped them to change their "processes," change they way they approach
each day, and manage their business and their daily activities?
"To me, that's what Grid is all about. It's about thinking differently.
You need to go back and redefine the problems you are trying to address
-- with an open mind toward the new possibilities that we're starting
to see with Grid and SOA."
"I think a big part of the problem here is in the way we label things.
We have too many people moving way too fast wanting to label the
solutions without taking time to articulate and fully understand the
problems they are trying to address."
He continued with, "Discussions of market size and available market give me heartburn."
We'll get back to Mr. Preferido, but first let us frame the problem I'm
hoping to address. Several other end-users I spoke with voiced their
frustration at trying to understand and model the market dynamics,
while several of the vendors, those with a stake in understanding the
landscape and available market, voiced their confusion and frustration
as coming from the way analysts are looking at the market.
Well, guess what folks. The analysts are every bit as frustrated.
This is a difficult job. We seem to have forgotten that the technology
sector lives in a vicious cycle. The end-users tell the analysts what
technology they need and what products they will spend money on. Then,
the analysts tell the vendors what demographics will be spending money
-- how much and on what. Then, the vendors use the information from the
analysts to validate the available market and make decisions on product
direction, engineering, manufacturing and timing. Then, the analysts
tell the world how things are going based on the progress of the
vendors and the sales and installations of the end-users.
So, imagine how difficult this is when everyone is speaking a different language.
That's pretty much where we are with trying to define the Grid market: square pegs, round holes, different languages.
Kelly Vizzini, chief marketing officer at DataSynapse, summarized it
pretty well. "The very fact that Grid computing is applicable across
the entire application, system and data stack guarantees that vendors
in each space will have their own messaging."
In my opinion, the industry champions have inadvertently created a
communications challenge and it has placed the Grid market in a cloud
of confusion. I also think we can better define the Grid market if we
step back -- or up -- and look at Grid within the context of the bigger
Experts such as Fogel and Vizzini, along with well-respected Grid
evangelist Wolfgang Gentzsch, a former Sun Grid strategist now with
MCNC, agree on a few things, though. The market opportunity for Grid
needs to be looked at within the context of its impact on future
enterprise infrastructures. Will companies continue to invest in new
computing platforms, software and networks without Grid? Sure they
will. However, let's not get hung up on trying to define exactly how
many dollars are being spent on Grid infrastructures. Instead, let's
focus on the benefits brought about by a well-planned Grid
If you take an existing computing infrastructure and make it
Grid-enabled, do you go back and reclassify all the money you spent
previously on that infrastructure as an investment in Grid? Of course
not. Making it Grid-enabled is adapting it to the times. Money spent to
buy servers or clusters prior to the implementation of a Grid is a
hardware purchase of servers or clusters and should not be counted as
an investment in Grid computing.
Now, back to Mr. Preferido.
We took a break to get some hot coffee and a pastry. Now it was time to get down to something more concrete.
Me: "I heard -- very confidentially of course -- that your organization
is planning a major computing system infrastructure procurement in 2006
that would also represent a major venture into building a Grid
computing environment. I heard that you were allocating close to $100
million to build this new Grid computing infrastructure."
This brought a big smile, a chuckle and a comment of, "Well, isn't that interesting?"
Mr. Preferido: "I know where that number came from and that's not exactly how I portrayed the information."
Using the base number of $100 million to build a new computing
infrastructure, we worked through his very rough plans and agreed that
the following would be the most accurate way to categorize how the
money was going to be used: $20 million in networking equipment and
upgrades; $60 million in new computing platforms, specifically clusters
(with software licenses) and a significant amount of storage;
$150,000-$200,000 dollars in Grid infrastructure middleware; and the
remaining $19 million was for facilities, staff and operations costs.
Now, just to recap, I was told in confidence by someone outside of this
organization who frequently states opinion about the potential for the
Grid market, that this company was planning an investment of $100
million to build a major Grid computing infrastructure.
Mr. Preferido said he would classify this as a $200,000 Grid sale, as
everything would be installed, up and running regardless of their
ability to Grid-enable the entire infrastructure.
Do you start to see where the disconnect is? At least one person out there is counting this as a $100 million Grid sale.
The only way to tell what's really going on is to follow the money.
I've seen several impressive marketing budgets in the Grid space, but
no one is talking about revenue. I've talked to numerous investors
about the space and I certainly don't have to be sold on the
excitement. But, is anyone really making money in this space?
Once again, I turned to DataSynapse's Vizzini, who told me the company
has experienced triple-digit revenue growth every year since its
inception. Triple digit revenue growth! That's impressive in any
market. And, sure enough -- after speaking with a number of end-users
-- if there's anyone making money out there, I would believe it is
What I find quite refreshing is that DataSynapse is one of the most
conservative companies out there when it comes to talking about the
Grid computing market. The DataSynapse spokespeople are very careful to
clarify, be factual and avoid hype. Bravo.
We discussed some more that may show up in a future column, but we had
to bring this interview to a close. I feel like I was able to make more
headway in understanding some of the Grid market dynamics just by
taking an hour to enjoy a good cigar, sip some very strong coffee and
have a non-hurried conversation where no one was trying to sell
Thanks, Mr. Preferido. I will be sending you that box of cigars as promised.
So, if we don't bog ourselves down in the mud with all the confusion
over labels, and if we focus more on the big picture where Grid is an
enabling component of a new enterprise architecture, and we stop trying
to create a market projection that's intended to pad an ambitious
marketing plan, we may just get through this.
The pegs do fit -- if you take the time to find the right holes. Maybe everyone is just trying to move a little too fast.
There are some great success stories out there. There are many
interesting examples of Grid being applied to boost productivity and
performance. And there's a tremendous opportunity to start thinking
about the way we approach computational and data manipulation problems.
Therein lies the challenge.
It's about a new way of thinking. Think about it.
And while you're thinking about it, relax. It's all good.
About Mike Bernhardt
Mike Bernhardt is a contributing writer and columnist for GRIDtoday
a consultant with 30 years of experience in helping companies develop
sound, executable take-to-market strategy. Bernhardt can be reached at firstname.lastname@example.org