Peeking Inside Cloud Computing

Peeking Inside Cloud Computing

Ed Sperling / forbes

Cloud computing may be the latest in a barrage of techno buzzwords,
but it’s already in widespread use across corporate enterprises and by
many consumers. Even more surprising: Most people who use cloud
computing don’t realize it.

Just how big will this phenomenon actually become, and is it really
new? More to the point, why should chief information officers care? To
help answer those questions, Forbes.com caught up with Peter Bell,
general partner at Highland Capital Partners and the former chief
executive of StorageNetworks.

Forbes.com: How new is the concept of cloud computing?

Peter Bell: It’s different than what we used to call
utility computing, and it represents a fundamental shift. It’s not just
moving a server from my data center to a hosting center and calling it
a "cloud." It’s a shift in how the whole platform and application stack
is built. That is when it becomes a different computing model. We’re
just beginning that. It’s not really there yet for enterprise class
computing, but there are companies we fund–startup companies–that
definitely are building their environment in a cloud. They don’t think
of it as outsourcing. They think of it as a much more scalable,
cost-effective way to do something that’s not core to a business but
which is critical to their business. It’s a big shift.

How real is the market for this?

There’s a lot of hype around it, and a lot of venture capital
has gone into it. There will be a lot of carnage. But there will be
some good companies that emerge, similar to Web 2.0. Was Web 2.0 a
buzzword? Yes. But when you look at Facebook or MySpace or Gmail,
there’s pretty significant value there. It will be the same thing in
the cloud computing space.

Is the upshot of this efficiency but with lower quality service?

In some ways, this is scary. If you have a problem with Gmail, whom
do you call? It’s one thing if you’re at home, but what if you
outsource your mail as a company? There’s no one to call. It’s very
different than calling your IT support or your storage vendor. That’s
why it’s going to take some time to figure out the value of the
computing solution, the service-level agreement, the hand-holding, the
expertise, the implementation–there are a lot of pieces.

If you have a cloud and it fails, there isn’t anyone to call. That
may be OK if you’re bootstrapping your operation, because you may be
able to recreate your infrastructure. But as you get into
mission-critical or more established business, there will need to be
multiple levels of solutions. There is some interest already in
application-specific clouds. Salesforce.com
(nyse:
CRM

news


people
)
is a good example. They may not be what you would describe as classic
cloud computing, but they picked CRM and they nailed it. It used to be
a multimillion-dollar enterprise license and three times that from a
systems integrator, and Salesforce.com made it so you can be up and
running for $100 a client in literally minutes. And they gave you
someone to call. Other folks will emerge with some application-specific
ability, and those will be the ones who get traction in large
enterprise-class computing.

What does this mean for security?

There are a number of areas of significant concern. There’s also the
whole perception problem. No one wants to admit they have in-house
security breaches, but the vast majority of security breaches occur
inside the firewall. I’ve seen companies outsource application
development to folks they may not know in a country they don’t have
much experience working with. That means a company with which they have
a complete arm-length relationship has access to the entire corporate
database.

That wasn’t called cloud computing. It was called outsourced
application development. In the cloud, I think companies that can
deliver a solution where they have a secure entry into the firewall
will get the initial traction. You’ll also see new companies emerge
that have optimized a security solution around the clouds. We’ll have
security issues that didn’t exist in a non-virtualized environment.
There are a lot of holes that will take time to be integrated tightly
with various levels of computing stack.

Isn’t the risk greater in cloud computing, though? You don’t really know where the computing is happening.

In some ways, not knowing is part of what you want. The more you
delve in, the more it’s like saying you want to go back into the
kitchen to see how everything is prepared before you eat it. It doesn’t
work. It takes too long to get your food, the restaurant wouldn’t make
any money and you might not like what you see.

If you use Gmail and Google
(nasdaq:
GOOG

news


people
)
calendar, you don’t need to know how they’re doing it and what their
storage model is. You trust them. If you have more sensitive
information, you handle it a different way. But at some point down the
road–and for some early adopters, it’s now–enterprises will be
comfortable trusting a third party with capabilities the enterprise
feels they have a handle on today. The model breaks down if you have to
know every vendor and every integration point.

Does that give big name vendors like IBM a leg up on competitors?

They probably have a different understanding of the enterprise
environment, both in terms of process and support. They may be better
from that standpoint, but they may not be better from a technology
capability and cost standpoint. You have to balance the IBM
(nyse:
IBM

news


people
)
support and trust with the innovation of a smaller company. What you’ll
see is that IBM will play a role, along with other large integrators.
Then you’ll have companies like EMC
(nyse:
EMC

news


people
)Oracle
(nasdaq:
ORCL

news


people
) and VMware
(nyse:
VMW

news


people
)
providing cloud computing. Those companies will emerge as cloud
computing providers. You’ll also see software companies emerge with
on-demand software, and new companies that provide some kind of value.
The ones I worry about are the ones that come to me and say they’re
going to have a cost advantage. If you’re a startup, you’re not going
to have a cost advantage over Amazon. Amazon has a better balance
sheet, more scale, operational proof in its ability to deliver. So for
startup companies, what do you offer that’s unique and what’s your
ability to partner with an IBM or EMC or Amazon?

Are startups offering something unique?

Yes. But in the product world, you have to ask whether it’s a
feature, a product or a business? Sometimes, with the subcontractor,
it’s a nice capability but can you build a business around it? How do
we as investors realize value down the road? You have to be able to
build a business. There’s a big difference between being a feature
versus being a business.

Where are the smart investment dollars going in the cloud computing area?

We’ve looked at a lot of cloud computing companies and tracked the
number of companies that have been funded, and it’s daunting. There are
too many companies looking for 10% of the market. What we’re looking
for are a couple things. There has to be enough value in the features
they offer or in their subcontractor capabilities that customers will
demand those features or capabilities. If management capabilities are
optimized for the cloud, then the customer will tell the likes of IBM
that they need to include it. So those capabilities may be part of the
solution from an IBM even though the customer doesn’t want to buy that
same technology from a startup.

When we do our due diligence on potential cloud investments, we may
call 30 to 40 respective customers and understand how important the
proposed capability is.

The second thing we look for is whether the team has demonstrated an
ability to create meaningful partnerships and alliances. You have to
have confidence that you can build a company and startup as an
entrepreneur.

Many of these companies are big-name brands that are part of
large ecosystems. There’s a whole other side of small startups looking
for the lowest price, right?

That’s right. I’ve made a couple trips to China. The opportunity for
new players exists because there isn’t as much legacy infrastructure.
At the same time, brand really matters there. It’s important from a
relationship standpoint that you have some big company saying this
company is OK.

New companies in China may provide more of the solution than they
might here. In China, there are a lot of nuances with firewalls. Data
often can’t leave the company. That means a cloud computing player has
to be local–and the cost to localize can be daunting. We’re starting
to see partnerships with hosting providers, telecommunications and
network providers in various foreign markets all as part of the
solution but leveraging the local providers for customer management

When do established companies buy in?

A lot of noise combined with a challenging economic environment plus
a lot of choice create long sales cycles. Those are killers for all
startups, and cloud computing companies are no exception.

What will jump-start the cloud computing ecosystem then? A single critical event? Or will we just one day wake up and notice it’s already happened?

It’s going to be the latter. I think of it like virtualization.
VMware was a company you didn’t hear about five years ago. Today, if
you call up any IT executive, they are virtualizing big chunks of their
environment because the value proposition is there on day one. That
didn’t happen overnight, but as VMware got closer to their IPO and you
realized how simple the idea is to understand, you could see this was
going to be a very important software company. Cloud computing is the
same because it makes so much sense. All these data centers do about
the same thing. You want five nines of availability, you want data
recovery in a relatively quick time period and you want high bandwidth.
Most large companies have the same things, and they have expensive
people managing them. That doesn’t make sense. There are reasonable
issues involving security, job loss
and loss of control, but cloud computing is going to happen. Then, if
you add in issues of power and cooling and Green computing, you now
have 20 data centers in an area where you can have one in a different
power grid and at a much more cost-effective rate.

Leave a Reply