The other day I responded to a tweet from James Urquhart that prompted an interesting discussion:

“Does a successful virtualization strategy take away from the ROI of cloud computing?” <- Interesting. Not enough to avoid cloud, me thinks.“

What that made me start to think about was the underlying value proposition and ROI of cloud computing and the juxtoposition of operational and financial goals.

Under Urquhart’s perspective, a company that has deployed a successful VM has a lower Cloud strategy ROI than a company that has not deployed VM.  This may be true from a raw number perspective because the company with the VM strategy is already getting much more benefit from their existing infrastructure.  But they may still fight problems with scalability, depending on the projected use of their application, which might push them towards the cloud.

But it brought me to a very practical question that few are talking about: “How can companies maximize the value from their existing IT infrastructure when planning a cloud strategy?” If the Cloud is all about ROI, how should companies factor in their existing investments?

Now most public cloud advocates talk about the Cloud as if there is no existing infrastructure.  All you hear is No CapEx, Lower OpEx, Unlimited Scalability, Superior Performance.  Yet most of the people they’re trying to sell to have hundreds of thousands, if not millions of dollars in infrastructure investments.  For those companies the ROI of the Cloud is a much more difficult proposition because it leaves them with dead assets.  Yes, they can try to re-purpose those assets to run other systems, but that would require some detailed analysis and planning of their entire infrastructure needs, look closely at when to retire servers or other equipment and build a phased plan to enter the Cloud.  Quite frankly that kind of analysis is probably long overdue anyways.  But in today’s economic environment – which is improving, yet still a long way from boom times – people are trying to figure out how to squeeze more value from what they already have, not how to abandon it more quickly.

And for those who are thinking about evaluating a private cloud strategy (although a cloud on your own infrastructure in my view is not a cloud), they may be able to get more value out of virtualization and other strategies, but they lose the operational efficiencies and instant scalability to match spikes in demand that the Cloud has to offer.  It’s an alternative to the public cloud, but in many cases still doesn’t solve the business problem that moving to the Cloud is supposed to address.  As an aside @krishnan wrote a good post on his own changing views regarding the debate between public and private clouds and helped spark a few ideas for this post.

So it leads me to this.  If “Cash for Clunkers” worked for the car industry (at least to the tune of 690K vehicles), why not bring the concept to the Cloud?  Why not “cash for Infrastructure”? When you think about it, it’s a virtual rush to acquire clients.  While there may not be true lock-in, it’s certainly a hassle to move operations from Cloud to Cloud, so it’s time to lock-up customers.  So why wouldn’t deep-pocked companies like Amazon, Google or Microsoft offer some sort of upfront cash for a company’s existing computing infrastructure assets to lock in a customer now?  Of course companies on the next tier like OpSource, Rackspace or GoGrid probably don’t have the resources to offer it, so they’re at a disadvantage.

Does anyone have the “onions” to make this offer?  I’m sure there are a lot of companies who’d be willing to take them up on it.  If any of the above companies run with my idea, I only want 0.5% of the assets they acquire in compensation

UPDATE: @lmacvittie had a great suggestion on Twitter to improve the idea. It doesn’t have to be cash. Even credits for cloud usage would be a good incentive. Maybe now guys like OpSource and Rackspace have a way to play.

Yesterday my CTO and I were given a demo of OpSource’s new OpSource Cloud, now in private beta. My company, Symphony Services, already is an OpSource partner, but as we’re starting to get more involved in developing Cloud-based solutions and POC’s, we figured to take a look under the hood. Admittedly we didn’t see much, but here are some early observations:

• Nice admin features: You start with a primary admin account which can then create any number of sub-admin roles with a lot of ability to set permissions based on role. Seems more granular control than what you have with Amazon.
• You have to set up a VLAN at the outset, so there’s an additional layer of security before the firewall.  Unless and until you open up the firewall the cloud is locked down.  Good fit for people who are looking for private clouds.
• The network set up seems superior in some ways to what AWS offers, particularly the IP mapping ability
• They offer a ‘local’ storage option as opposed to just cloud storage, which can help with application performance on data-intensive operations.  But today the limit is 400GB per virtual server which seems kinda limited, especially when you consider how many people have terabyte drives in their homes now and that many of the really large information services providers have petabytes of data that they’re working with.  But then again, OpSource seems to be focusing on the Enterprise rather than large service providers, so it may not be as much of an issue.
• While spinning up new servers seemed very simple and there was quite a lot of control given, what we didn’t see was anything about dynamic provisioning.  Perhaps we’ll see that once we start playing around in the sandbox during the private beta.
• With their focus on the Enterprise they’re making some clever use of their OpSource Billing solution (that they acquired when they purchased LaCayla about a year and a half ago) by being able to bill at the sub-account level, ostensibly to help charge different departments based on usage.
• The UI is a bit primitive, but they copped to that right up front and there’s no reason to believe that they won’t have this fixed when the product is ready to go live.

So we certainly saw enough that we’re going to give it a test drive and become part of the private beta, and perhaps run a few client POC through it as an alternative to Amazon Web Services.  Keep you posted as we learn more.