Cloud Computing and Evolution; Monkey Revenge

There’s a ZDnet blog by Mary-Jo Foley titled “‘Pri­vate cloud’ = just another buzz­word for on-premise dat­a­cen­ter?” I caught the post on twit­ter (thanks Mr. Linthicum!), quickly favor­ited it because I knew I’d have a strong opin­ion on this one!

Turns out the arti­cle wasn’t what I thought. The arti­cle was about Microsoft’s odd dual-approach cloud strat­egy, a ‘pub­lic’ cloud via Azure or a pri­vate cloud via ‘other tools’ that are not Azure. Frankly, I think that’s ridicu­lous, unless those other tools are the same ones Microsoft uses to run the Azure cloud, but… I’d like to talk about Mary-Jo’s orig­i­nal topic.

You see, I think for many, ‘pri­vate cloud’ will unfor­tu­nately be a fancy of way of talk­ing about their dat­a­cen­ter. I think the many will be wrong.

On tech­ni­cal archi­tec­ture draw­ings, a dat­a­cen­ter may look like a cloud… it may have all the hard­ware, net­works, and even server vir­tu­al­iza­tion and fea­tures of a cloud, but I strongly sus­pect that most IT orga­ni­za­tions will stop there, and not think about the results achieved, and there­fore not go far enough with their dat­a­cen­ters to make them true clouds.

What’s the dif­fer­ence? Well, it’s really in the “feel” the “intent” and ulti­mately in the results.

As a result of my deep inter­est in lead­er­ship, and some inter­nal projects I’ve worked on in the past, I’ve become a stu­dent of social com­put­ing. I think most peo­ple relate me to enter­prise mes­sag­ing archi­tec­ture, and less so to social com­put­ing. That’s OK, but I’ve real­ized that I really get social com­put­ing, per­haps because I’m an innate leader, and social com­put­ing is about lead­er­ship. (more on that in another post another time)

My point is, you can have a list of “check boxes” to sup­port on a web­site and think you’ve built a com­mu­nity. You’re stumped when you’re not get­ting “com­mu­nity results” and can’t fig­ure out why. It’s because the feel of the inter­ac­tion is forced, and the intent was to build some­thing viral… yet if it’s forced, peo­ple won’t be ener­gized, and it won’t suc­ceed. An exam­ple on one “com­mu­nity” I’ve been exposed to… I was adamant that it sup­ported tag­ging as a way to cat­e­go­rize infor­ma­tion rather than a fixed tax­on­omy based upon what some archi­tect (who’s never been in the role of the users) thinks is appro­pri­ate. I was promised it would sup­port tag­ging, and it does… for the peo­ple who cre­ate con­tent. User tag­ging is not allowed. Duh? What? Yeah, user tag­ging not allowed. If users can’t par­tic­i­pate, it’s not a com­mu­nity, but that noth­ing to do with the tech­nol­ogy employed (the tech­nol­ogy archi­tec­ture sup­ports it).

So, list of fea­tures. Check. Intent. Fail. Fail big. Project fail. (That’s a pre­dic­tion folks… it’s a project that hasn’t fig­ured out that it’s failed yet… don’t tell them.)

Same goes for cloud com­put­ing. You need to imple­ment tech­nol­ogy… and intent. Tech­nol­ogy alone isn’t enough.

So, what’s the intent behind cloud com­put­ing? I’ve got a few “mis­sion state­ments” to describe it, but one I like that’s catchy, is:

Turn­ing a com­mod­ity into a utility

Unfor­tu­nately, most IT orga­ni­za­tions (and in fact, the whole ITIL method­ol­ogy) focus on “projects” not “prod­ucts” and until IT approaches the cloud as a “prod­uct to evolve over time” they sim­ply won’t be suc­cess­ful. You see, it’s never “done”, and until IT under­stands that we’re just build­ing cas­tles in the sand below the tide line.

Finally, I hope ven­dors will ship tools and prod­ucts that help com­pa­nies turn their dat­a­cen­ters into clouds, I know of big and small com­pa­nies who are (attempt­ing) to do so. I wish them luck, but am skeptical.

It’s a long evo­lu­tion from dat­a­cen­ter to cloud. Remem­ber the last time we evolved and what hap­pened to the mon­keys? I think this time the mon­keys real­ize they’re endan­gered, and are not going to go down with­out a fight.