The 451 Group is brimming with personality. I know, that's kind of a weird commentary about an IT analyst firm, but they, of all the groups I interact with in the IT operations space, don't check their quirks, sense of humor, and appreciation for irony at the door. And that's a good thing.
In reading their recently published year-end analyses and predictions about cloud computing, eco-efficient IT, and virtualization, however, I think they ran into something that forced them to keep some of their trademark commentary in check: there was a lot of stuff to write about.
Anyone watching the emergence of the cloud computing market in the past year has to feel a bit out of breath. The speed with which things appeared, changed, and changed again, was stunning ("a faster rise than the Web itself," notes the 451 Group's William Fellows and Antonio Piraino). In fact, their 2008 wrap-up and 2009 look forward, as a result, actually read a bit more like a list of the many vendors and trends that entered (or will) enter the scene. Not much space for pithy comments, though they managed to squeeze in a few nonetheless. (Besides, they have the rest of the year for that, right?)
Between gulps of air, Fellows and Piraino did manage to sift through the many pieces parts and articulate some interesting tidbits and trends, though. I'll highlight a couple interesting ones here.
An internal cloud is part of a continuum, which includes grid and utility computing. A lot of the conversations in 2008 were about what all this cloud stuff actually is. The 451 Group noted that internal clouds were one of the things they hadn't predicted to appear in 2008, but had shown up just the same. They called cloud for enterprises "grid done right" (right before pronouncing "grid is dead"). This perspective doesn't surprise me, given their strong background in following the grid computing market.
They did catch the subtle differences between grid and cloud computing: "Where grid concerns a fixed resource pool, the cloud implies a flexible pool; grid provisions capacity where the cloud provisions services; grid is science and the cloud is business." They saw a sweet spot of activity from their customer base (which I know has a good sampling from the huge financial firms for starters) in "the creation of shared internal resources, or clouds, whether or not they are sold internally as that." They even suggest selling an internal cloud project under some other name if the hype is just too thick for your organization to handle. Bottom line is that in 2009, "[c]reating internal clouds is now on the agenda for early adopters."
"Cloud is a way of using technology, not a technology in itself…and economic imperatives will drive it." Fellows and Piraino noted in their 2009 cloud preview that virtualization and automation are change agents in all this (that virtualization comment echoes sentiments from IDC's Al Gillen). "Accelerated consolidation, virtualization and automation strategies will be the building blocks of shared infrastructures," they say, "delivered as a self-service utility with retail discipline. This internal cloud is the logical end point of these combined activities." And, they added, the recessionary forces at work mean that "the opportunity to move spending from cap ex to op ex can't be underestimated."
Here's the good news I read into that, from another part of that same report: the cultural and organizational barriers that have been the biggest inhibitors for adopting shared infrastructures may have finally met their match. The 451 Group says that the recession, painful as it is, could actually be the thing that gets these barriers to fall. (It only took a global financial meltdown -- is that saying something about our industry?) Companies will do whatever they need to in order to "weather the current storm." Says Fellows in a different report on this same topic: "given the uncertain conditions and focus on cost, we think the cloud computing economic model will win out."
As for 2009 predictions, what's their tip for beating the recession/credit crunch? "Start to use cloud computing." Pretty clear, if you ask me.
Key challenge for 2009: how do you "bottle the magic" of providers like Amazon and Google? The key challenge enumerated in all of this, though, is to "bottle the magic" that makes folks like Amazon and Google "so cost effective and so elastic" and apply it to any type of cloud implementation -- inside or outside your data center. Now, Gartner talks about how you actually can run your IT systems for half what Amazon can do for you with EC2. Fellows and Piraino sound like extreme skeptics on that point: "The question is whether IT groups can deliver. ...Is it cheaper to do it in-house? Are existing management tools adequate? No."
Now, they missed pointing out that even "the magic" has some pretty big flaws. It's cost effective, yes, for the applications that can live within their constraints. It's elastic in a way not seen before in how companies own and use compute resources, but it is still very coarse-grained in what/how/when users can change. It's still waiting for a move toward policy-based automation to control the compute resources, enabling much less manual effort.
The 451 Group's further comments on a do-it-yourself approach are typically and justifiably what I'd call "skeptical until proven otherwise." For cloud control, they say, don't bet on things suddenly being less complex. "Virtualization and fungible resources introduce new management requirements out of reach of most of today's tools."
So what tools might those be? Well, they predicted that vendors like Scalent, DynamicOps, Q-Layer, and other private cloud enablers are likely to continue to get traction and attention. (The last company they named has certainly wasted no time.) In a separate report focusing on Cassatt's use of policy-based automation to support internal cloud computing, Fellows notes something that might help us stand out from this crowd: "unlike other vendors, [Cassatt] uses existing infrastructure, doesn't alter existing software, doesn't introduce new layers or agents." We certainly think those items are key, especially with rough economic waters threatening every organization from all sides.
(As an aside, Fellows did get a chance to use his fondness to point out market ironies in that Cassatt report. "It's not a little ironic," he noted, "that Cassatt's entire value proposition since it was formed has been targeted more or less on the anticipated benefits of internal cloud computing -- although, it wasn't called cloud computing until recently.")
What’s next? The "holy grail" is federating internal and external clouds. Fellows and Piraino set up a pretty big challenge for 2009 as well. "…[T]he Holy Grail of cloud computing may very well be the ability to seamlessly bridge both private clouds (datacenters) and remote cloud resources like Amazon...(EC2) in a secure and efficient manner." Their summary on this? "It will take time." Yup. No argument there. Lots of steps are being taken, but lots of work to do still.
So you can track the source material down (if you're a 451 Group client), below I've listed and linked to the three reports referred to in this post. The 451 folks also do a cloud blog you should check out.
- 2008 review -- Cloud computing, William Fellows and Antonio Piraino, Dec. 8, 2008
- 2009 preview -- Cloud computing, William Fellows and Antonio Piraino, Dec. 19, 2008
- For Cassatt, the cloud is the next driver of datacenter efficiency, William Fellows, Nov. 18, 2008