« Cloud Parsing | Main | Watching Microsoft Positioning its V12N Offers »
Sunday
Sep072008

Cloudonomics -- and its Ten Laws

Gigaom has a put out a piece by Joe Weinman, a Solutions Sales VP at AT&T Global Business Services. I enjoyed the laws, and I particularly liked Laws #2, #3 and #8.

I do, however, take some exception to the sentiment at the beginning of the piece. It feels as though the argument discussion begins with a presupposition: there's "good" cloud computing and "bad" (or at least, badly executed) cloud computing ... and it's all based on whether you use a utility service or decide to run your in-house datacenter as though it were a utility. There are advocates for either side, and they seem all too willing to take after one another with axes.

I'm not convinced there needs to be a dust-up about public utility clouds and private clouds. They have their respective advantages and disadvantages. What seems to be missing from the argument is the notion of spanning or cloudbursting. There's every reason to think that the benefits of cloud computing / utility computing are available to the organization that runs its "cloud" in a corporate datacenter, but makes use of the utility cloud when and if it's needed. This could be for scaling out an application or scaling it up (to use functionality unavailable within the corporate datacenter) for so long as the resource or functionality is necessary.

So why (other than for sales positioning) does the end-user need to pick one or the other? It seems logical (and practical) that for many enterprises, the hybrid approach will be necessary -- some while in transition, others permanently.

Let's stipulate that both the public compute utiltiy and the private enterprise cloud have their merits. And, furthermore, that the appropriate recipe for a specific enterprise may require a mix of both.

There, now. That wasn't so hard, was it?

The 10 Laws of Cloudonomics - GigaOM

Public utility cloud services differ from traditional data center environments — and private enterprise clouds — in three fundamental ways. First, they provide true on-demand services, by multiplexing demand from numerous enterprises into a common pool of dynamically allocated resources. Second, large cloud providers operate at a scale much greater than even the largest private enterprises. Third, while enterprise data centers are naturally driven to reduce cost via consolidation and concentration, clouds — whether content, application or infrastructure — benefit from dispersion. These three key differences in turn enable the sustainable strategic competitive advantage of clouds through what I’ll call the 10 Laws of Cloudonomics. ...

Reader Comments (2)

Rich,

You are completely correct. The emerging architecture, which I have been calling Architecture 3.0, is not pure enterprise data center or cloud: each has value. Per Law 1, the utility premium is only worth paying if the peak to average ratio is higher -- in other words, for relatively flat demand, it's better to do it in the enterprise data center. For relatively spiky demand, though, it's better to do it in the cloud. My utility computing simulation at ComplexModels.com demonstrates exactly where the breakeven point is for any demand type, but in general, the economically optimum point is to use owned, dedicated resources to handle the baseline demand (and possibly a little bit more) and then cloudburst for spikes.

Non-interactive, non-spiky demand can be handled in a centralized enterprise data center. However, per Law 8, I believe that as applications become even more interactive and content becomes even richer, a geographically dispersed architecture is essential, and most enterprises are consolidating data centers. So, to your point, a hybrid that uses an enterprise core for transactions of record, surrounded by a dispersed cloud-service provider edge for interactivity balances things nicely.
Sep 7, 2008 at 10:37PM | Unregistered CommenterJoe Weinman
Joe,

Thank you very much for the comment. I agree that the practical conclusion is (as we're both stating) a hybrid. Further, its the nature of the requirements placed by the application, mediated by the constraints represented by SLAs and compliance guidelines that will result in a set of solutions that satisfactorily performs for the end-user.

Because no single vendor of technology or cloud services has yet to deliver working, usable cloudbursting for production computing, I'm afraid we're going to continue to hear more debate about public cloud vs private cloud -- as though the two were mutually exclusive. If this conversation does anything to reduce the waste of blog space dedicated to that specious argument, I'll consider it a "win."

Finally, thank you VERY much for pointing out to me your ComplexModels.com site. Two thumbs up!
Sep 8, 2008 at 1:03AM | Unregistered CommenterRich Miller

PostPost a New Comment

Enter your information below to add a new comment.

My response is on my own website »
Author Email (optional):
Author URL (optional):
Post:
 
All HTML will be escaped. Hyperlinks will be created for URLs automatically.