CAMBRIDGE, Mass.--Cloud-computing events are in vogue these days. Wednesday's CloudCamp Boston (actually held across the river at Microsoft's Research and Development facility in Cambridge, Mass.) was a good one.
To be sure, I read a few online commentators who were of the opinion that the material in the formal part of the event--CloudCamp is organized as a combination of pre-organized talks and an unconference format--was far too basic. However, a lot of the questions I heard and conversations I had at the event suggest to me that a lot of people are still trying to get their heads around the basic concept of cloud computing. As a result, it behooves those of us who work with this stuff on a daily basis to remember that not everyone is quite so immersed.
Any event of this sort inevitably has lots of different simultaneous threads going on. However, here are a few that I think are worth highlighting:
Is security really an issue? The security aspects of cloud computing are often presented as a matter of trust in service providers and getting comfortable with the loss of direct hands-on control. Those are part of it certainly. But a number of discussions made it clear to me that the situation is a lot more complicated than developing a "comfort level" with the technology. It also goes beyond point products such as data encryption.
Christofer Hoff (who was in one of the unconference sessions I participated in) wrote a post a few days ago that gives a nice window into some of the complexities here. This isn't to say that security control and compliance concerns are stumbling blocks for moving all applications into a cloud, but they do have to be taken into account (in all their myriad complexity) for many core business applications.
What is cloud interoperability? Interoperability seems to be emerging as a bit of a contentious topic in cloud computing. This is partly because interoperability, like cloud computing itself, means different things to different people. Just about everyone agrees that base-level data portability (download your customer records from a CRM system in a readable format) is a must and that a nirvana of totally transparent computing delivery across providers is years away.
The confusion and argument comes in the vast middle ground.
A couple of things that people at the event seemed to generally agree with:
The higher you go up the stack from Infrastructure-as-a-Service (a la Amazon Web Services), the less interoperability you can or should expect. (And, indeed, the less interoperability IT folks are even necessarily looking for.) In other words, don't expect to transparently switch from one business intelligence application to another even if you can move and transform your data relatively easily.
A lot of interoperability will be provided not by the service providers themselves (who after all don't have a particular incentive to make switching away from their service easy) but by third-party "service brokers." We see a good early example of this sort of thing is RightScale who is essentially mapping the APIs from different IaaS (infrastructure as a service) vendors like Amazon and Rackspace to common interfaces in their management software.
Where's the business opportunity? Finally, in a session led by fellow analyst Judith Hurwitz, we discussed the opportunities for start-ups in cloud computing. One way to look at the opportunity is to break it down into two broad categories: companies and products such as development or management tooling that make clouds work better and those that are enabled by clouds. The latter includes software as a service broadly, but it could also refer to things like distributed testing that make use of geographically distributed public cloud providers.
As is the case with the software industry more broadly, there was general agreement that industry-specific products and services often offer better opportunities than big horizontal plays. For example, offerings that deal with data governance issues related to a specific vertical such as the pharmaceutical industry.
One thing there was consensus about is that setting up capital-intensive cloud infrastructure is not typically going to be a good start-up play. (In general, I conclude that far more start-ups are going to be enabled by cloud computing--vis the explosion of mobile applications--than will be involved in creating the platforms that support it.)
Cloud computing covers a lot of territory and there are still enormous differences in how familiar different people and organizations are with the particulars. However, one of my takeaways from this event is that we're seeing reasonable consensus on the basic taxonomy of cloud computing (even if we argue about terminology like "private cloud"). But, like peeling an onion, now that we're winding down some definitional and basic value debates, we're starting to get into more detailed and thornier questions about the particulars.