John Foley blogs about an interesting Gartner report that gives cloud computing up to 2015 for "mainstream enterprise adoption." He describes it as "a surprisingly conservative forecast for business adoption of cloud computing services." I actually thought the timelines were quite rapid and that we're just not sure about what Gartner means by "mainstream."
The Gartner report misses a lot of the interesting enterprise movement that got rolling in 2008. Smart enterprises are experimenting with the cloud today. They are cherry-picking apps that are appropriate for today's cloud offerings so they have the experience necessary to move more involved and sensitive ones tomorrow.
But one step at a time. Let's try the timeline from the Gartner report on for size:
2009 is for "pioneers and trailblazers" - I can't argue that enterprise adoption will go beyond trailblazers this year without having multiple vendors offer clouds with solid security certifications that enterprises can insert into their sarbox, PCI, HIPAA, etc. audits. There are a lot of really interesting, solid, cost-saving, time-saving things enterprises can do in the cloud in 2009, but they have to cherry-pick the right projects. Will we get to these security certs for cloud offerings and all that in 2009? Probably, but not early enough to move the mainstream needle.
Beginning in 2010 "cloud computing will appeal to a broader range of companies, resulting in a more mainstream user base" - I predict that it will take until next year for enterprises to have the red carpet rolled out in front of them leading them to the finish line with all the i's dotted and t's crossed. From books about cloud computing, to cloud consultants, success stories, ROI calculations, security certifications, enterprise offerings, yadda yadda. This is a lot of moving parts, most of which are not about whether EC2 works and is secure but about leading the horse to the trough.
It's between 2012 and 2015 that "cloud services reach mainstream critical mass and commoditization" - How long it is going to take AT&T, EDS, IBM Global Services, Savvis, and everyone else to have an enterprise cloud offering in full production, with all the services and consultants it takes to reel in the "mainstream"? Can I picture earlier than 2012? Mhhh.
The Gartner report misses a lot of the interesting action that has started to get into full swing. The moment sometime between 2012 and 2015 when the last enterprise moves into the cloud may be interesting, but it's far more interesting when significant movement into the cloud starts to occur. And that's definitely 2009 and more deeply so 2010.
Large companies don't run one app. They aren't sitting there waiting to decide when to move that app into the cloud. They have many apps. They all have different purposes, different stakeholders, different risk profiles, different security requirements, and different development timescales. The first ones could be moved into the cloud in 2008; more can be moved this year; yet more next year; and the last ones sometime in 2012-2015 if we believe Gartner. Great. Let's focus on those that can move now!
We see many interesting examples. There are marketing sites that don't have much security attached, are often under tremendous last-minute time pressure to add compute resources, and whose production is often outsourced. Cloud this!
Or batch computations that operate on public data using public algorithms. Sounds impossible to find? Well, a major pharmaceutic company has a good deal of those. The next step up is massive compute jobs that use proprietary algorithms that are not super-sensitive and operate on public data. We're working on some interesting architecture to enable these in the cloud. They have more steps leading to super secret data and very sensitive algorithms and we pushed those to the end of the queue. We'll get there, perhaps not in 2009 or maybe even 2010, although cloud time seems to be moving even faster than Internet time.
I could go on, but the point is that it will take enterprise IT shops a long time to build the internal know-how about the cloud to the point where they will be comfortable moving more and more mission-critical applications over to reap the benefits. Now is the time to start building the experience and understanding of what works and what doesn't. The smart ones clearly have started.