Cloud brokering isn’t about commoditization

One downside of the term "cloud computing" being rendered meaningless through overuse is that intelligent people can conduct two entirely different conversations and think they're talking to each other about the same thing.  The confusion is now spreading to discussions of cloud brokerages.  

Gartner's early (2009) use of the term referred to systems integrators and other third parties who would help their customers negotiate relationships among a variety of cloud vendors.  But recent buzz has refashioned the "broker" concept as a sort of commodities trader, aggregating raw computing power and creating spot markets.  It may be intellectually interesting to consider the feasibility of multiple public cloud infrastructure providers standardizing to the point where computing power could flow, frictionlessly, between data centers.  And the OpenStack initiative would be a good prerequisite to achieving this sort of liquidity.  

But suddenly we've started talking about infrastructure-as-a-service (IaaS).  There's nothing wrong with IaaS - I'm a big fan, and my company's products all run on Amazon Web Services - but to me IaaS is the least interesting layer of the cloud computing stack, at least from an enterprise business perspective.

IaaS is a "better than nothing" sort of proposition.  If you're going to need hardware, you may as well procure it via IaaS -- it's certainly preferable, in most instances, to expanding your own data centers.  But it's only a small step in the general direction of the future of computing, in the same way that "private cloud" (I can't believe I'm using that term in my blog!) is, well, better than NOT virtualizing across multiple private servers.  

Let's be clear, though -- software-as-a-service (SaaS) and platform-as-a-service (PaaS) are where the action will be at over the long term. You need to go "up stack" to SaaS and PaaS in order to realize many of the software benefits (in the end, that's why we have infrastructure, right? to run software?) that we public cloud computing advocates hold so dear:

• Faster vendor innovation, by reducing the need to support multiple instances, versions, and environment permutations.
• More agile software application development, by abstracting more low-level details away from the individual programmer and simplifying the build and deployment processes.
• More customer-realized business benefits, through automatic upgrades and backwards compatibility.
• More productive developers, by making it easy to spin up new, fully functional environments.

All of these benefits accrue at the software level, and you lose them all if you continue to build and maintain the "same old software" (single-tenant) and merely deploy that software to a multi-tenant IaaS platform.

I'd be happy to see "IaaS brokers" emerge, who can add efficiency to the market for raw computing power.  But a more fundamental shift in how businesses consume computing is coming in the form of true "cloud brokerages" that play counselor, negotiator, and traffic cop on behalf of enterprises, making multiple SaaS and PaaS platforms play together nicely.

We're moving into a new phase of maturity for cloud computing - shifting from departmental to enterprise-wide rollouts, from ad-hoc development to controlled migrations to production, from unintegrated silos to a broad cloud-based IT strategy.

With these added complexities, the stakes are growing higher for CIOs to properly manage cloud applications and platforms.  The "what do we have to lose" phase is coming to an end, and the "this stuff is becoming mission-critical" phase is upon us.

Glenn Weinstein is the CTO and co-founder of Appirio, where he oversees the CloudWorks and Cloud Management Center product lines as well as internal IT.