Nov
05
2010

Cloudy with a Chance of Meatballs

Sorry, could not resist. Having watched this wonderful movie and now writing about clouds is just too much. Excuse yours truly.

Back to work. Just the other day I read an interesting blog entry of Sandy Kemsley about IBMs’s cloud strategy. Here it is: Sandy’s blog entry

In a nutshell: she thinks that IBM uses the “cloud” name much too broadly, and thinks that most of the so called “private cloud” is in fact hosting or managed services.

Well, of course the buzzword of the year is always overused. In former times we called some of the stuff hosting, then ASP, then SaaS, and now we are talking about the Cloud. But essentially its’ the same deal, a method to solve the main issues or pain points of customers in the field:

1) high initial license and hardware investment

2) high TCO (administration, updates)

3) up front investment w/o any way to “breathe”

IMHO any service oriented environment can be qualified as a “cloud”, alas a private one, if it solves all of these points. Because that is something that you cannot do with simple hosting or managed services. We are talking about hardware AND software here. And no up front licenses. Pay as you go. And pay more as you profit more.

I strongly believe that a lot of our customers will move to a SaaS based offering within the next 2-3 years. Will there be public clouds for them? Most likely not. Public clouds more or less address new customers for new solutions. Typically without too much of customer specific adaptation.

Existing solutions will in most cases need a private cloud, and will not be suitable for a public cloud. Reasons which come to mind are:

- integration with customers environment (talk about SAP, Fax servers et al)

- highly customized applications

- sensitive data which must not leave the customers premises

- sensitive data traffic which must not leave the customers network

In all cases, the key to success is scaling and monitoring. A private cloud still must address the issue of TCO, so it must not create costs by having non-standard platforms or being in need of constant supervision. And it must be scaling easily, which is best done by using a swarm of VM clusters, which can be scaled on demand.

Once this is available, the work setting up such a private cloud for a customer is mainly by moving the customers application to the standardized environment and setting up the monitoring (for standard and application specific areas). So the customer has a low ramp-up and can profit from the removed TCO costs immediately. In addition, he will probably save by moving away from license based maintenance to a transaction/storage/active user based pay-per-use.

All in all, just reducing the “Cloud” topic to the well know Public Cloud area will remove a big part of the potential market and lose a lot of punch for the cloud idea. So if you want meatballs, you need all clouds you can get.

(from: http://thescepticaltechnologist.blogspot.com)

Post to Twitter Post to Delicious Post to Digg Post to Facebook Post to MySpace

Download PDF
Written by Axel Kroll in: english |

No Comments »

RSS feed for comments on this post. TrackBack URL

Leave a comment


× one = 1

Theme: TheBuckmaker.com Web Templates | Bankwechsel Umschuldung, Iplexx IT Solutions