Clouds are nothing new....

If you look at the most recent Gartner hype cycle for cloud computing it shows that we’ve just passed the Peak of Inflated Expectations and are headed for the Trough of Disillusionment. This paints a picture that cloud computing, and cloud services for that matter, are relatively new – or at least the “concept” is new.

I’ve probably had a thousand discussions on “the cloud” with my colleagues and associates. These have ranged from architectural to philosophical, but in the majority of the discussions the notion of when cloud computing first appeared on the landscape always seems to spark quite a bit of debate. I’ve always been of the opinion/position that cloud computing is not a new concept, and it’s actually much older than we think. Notice I said “concept”. I want to clearly distinguish this from capability, which has dramatically advanced over the past few years.

I just finished watching Derrick Harris’ interview with NetApp founder, Dave Hitz at GigaOM’s Structure 2011 event. Dave has some pretty interesting perspectives on the cloud computing model, some of which I agree with and others, well maybe not so much. One thing that I do strongly agree with him on is that we as the consumers of cloud computing are to a large degree confused about what clouds are or are not. This, in my view is due in large part to the industry itself, but also to another large degree in our belief that cloud computing is some mystical or magical capability that solves all problems. If we stop and think about it, the notion of consuming services that are provided elsewhere is nothing new. Let’s step back in time for a moment and look at earlier versions of this model.

Dave uses the example of processing his assistant’s time card to show that stuff already resides in the cloud, and those companies that don’t believe they are using the cloud are either misguided in their understanding of cloud computing or are in denial. I like Dave’s example in that it reinforces my position that cloud computing and services have been around for a long time. If you agree with his definition of cloud computing then in essence cloud computing has been around since the early days of IT outsourcing – the business which EDS launched in 1962. Granted the “capability” was vastly different, but the essential elements of the business model were the same. Customers paid somebody else to provide services for which they consumed on a “per sip” basis. Take ADP payroll services as another example. For years you simply mailed, faxed, emailed or uploaded your raw payroll data to ADP and they returned checks. Again, a simple form of outsourcing and “cloud computing”. As a company I didn’t have to buy the payroll system and associated computers and storage to process the weekly payroll. Hopefully you can see where this is going.

I can let someone else buy the computers, someone else buy the apps and I access it over the internet and pay for it by the sip. That’s my business definition.” – Dave Hitz, GigaOM Structure 2011

An interesting point that Dave makes from Clayton Christensen’s Innovator’s Dilemma is that in general technology gets better faster than user requirements increase. For the most part I agree with his perspective. I think that is what has happened with cloud computing. Technology (e.g., virtualization, internet connectivity and the “network effect”, socialization to name a few) has advanced so quickly in the past 5 years that it has outpaced our ability to build a suitable business model around its “mass consumerization”. Thus, we have lots of different models for cloud computing and services. Later on in the video Dave talks about storage virtualization and cloud storage. Technologically we have great capabilities, but have yet to define the right mix of services to consume what the technology offers. I think Dave offers a great perspective on how and where cloud storage will make sense – not as a distinct, standalone capability where I place all my data in isolation, but as an integrated solution with the application that enhances cloud computing performance.

As I noted above, I agree “for the most part” with his perspective. Here’s an example of where I think the reverse of the Innovator’s Dilemma is true. In 2005 I led the design of EDS’ “utility infrastructure” architecture. In simpler terms – a model for a shared (multi-tenant) computing, networking and storage services that could be delivered onsite for a customer or from a offsite location and accessed remotely. The resulting 47 page architecture definition titled Public and Private Utility Infrastructures outlined what most clouds look like today. But there were some key elements of the architecture that weren’t yet mature enough for enterprises to buy into the model. Virtualization, while out of its “infancy” stage was not yet widely accepted. The internet had not yet been embraced as an acceptable method of access for most applications, and security (while still a big concern today) was a major roadblock to adopting a shared infrastructure model. And then there were nagging issues like licensing and such… But when you step back and look at it, we basically offered cloud computing. We just didn’t call it that. So again, to my point, clouds (or at least the concept) have been been around for a while.

Another interesting point that Dave makes is who he thinks will be the cloud providers of the future – stating that those who are doing “weird little consumer stuff” may be the ones that surprise us (e.g. Zynga). I think this somewhat reinforces the notion that “clouds are what we want them to be” and there is no industry definition that neatly overlays the model and gives everyone that “aha” moment. I take a somewhat more pragmatic approach to it and believe that the industry will evolve to a crisper service definition, much like IT outsourcing – which took quite a few years to mature to the point where there was an ecosystem (vendors, licensing models, third party intermediaries, etc.) to support it. In the meantime, I see a lot of hope on the horizon in efforts like OpenStack and the Open Data Center Alliance which are starting to deliver some pretty good open source solutions and use cases that we can leverage and hopefully start creating a common understanding and service definition for this thing called “the cloud”. In the meantime, enjoy the ride!

------

For this and other topics on the digital age, visit my blog at www.robertkeahey.com