Wednesday, March 7, 2012

Is The Cloud Prepping for Hockey Stick Growth?

(Updated 3/10/2012)  
There are a number of significant things at play that may indicate the start of a hockey stick growth spurt in cloud computing.  Some may think it’s just a price war but I don’t think so.  In my last article I covered a number of changes that Amazon has made to their AWS offering.  In just the last couple of weeks, Amazon has started lowering their prices on various components of their AWS portfolio including S3 which saw a 192% year over year growth in 2011, and EC2.  This week, Google and Microsoft announced lower prices for storage.  In addition to this, recently announced pretty good first quarter results, highlighting two big wins, HP and Time Warner (Per ZD Net Article).  It’s hard to tell but may also be lowering prices a tad.

First, let me cover the topic of cost reductions that we are starting to see.  If these are related to a step function due to the economies of scale for storage and compute capabilities and not pure price wars (although there may be some price wars going on), this is a natural outcome as we shift from local compute and storage to the utility cloud model.  In Nicholas Carr’s book, “The Big Switch”, he described a similar situation during the birth of centralized power generation.  The cost to consumers to buy power goes down as more customers start using the centralized service who benefit from economies of scale and better asset utilization.   If this is indeed what is happening, the growth in cloud computing will only accelerate as the ROI increases for new customers.  

Software as a service (SaaS) has been around for quite some time and a lot of companies have been using it in one form or another.  The news that HP and Time Warner are switching to (SFDC) is coming sooner than I expected and this is big news.  Fortune 100 companies like these two have pretty conservative IT shops and the willingness to switch completely from on-prem company run assets to the cloud is a big deal.  I think it’s a tectonic shift.  HP has a large sales force (possibly 35,000 according to TMC) so this is big bet on SaaS.  According to a TMC article, (along with others), HP’s switch from Siebel to SFDC was rumored as far back as November 2010.  SFDC has other big clients but pulling Fortune 100 customers from Oracle is significant.  This announcement helps SFDC gain credibility, showing that big IT shops that are normally resistant to moving  out of their data center, feel comfortable with security, reliability and other factors to jump on board.

The next big indicator may be how much growth in platform as a service (SaaS) and infrastructure as a service (IaaS) from Fortune 500 companies.  SalesForce platform as a service,, is a natural segue for their big customers to try PaaS.  SFDC abstracts out a lot of the underlying services (like database) from the application and makes it easy to integrate with other SFDC products, like chatter, CRM, Knowledge Center and more.  If companies are going to build additional sales tools for their sales force, why not use SFDC PaaS to do it given the ability to integrate all of your sales tools under one umbrella and general UI.

Time will tell, but my prediction is that the Fortune 500 will continue the march toward SaaS and start to experiment with PaaS in 2012.


- Chris Claborne


  1. Did you see this I-Week analysis that totally contradicts the "savings" of AWS?

    Interesting reading, but I thought somewhat flawed.

    1. Thanks for the pointer. I've seen several cost analysis done on IaaS, and many have shown that it's more expensive than on-prem.

      One area that everyone (writers of geek zines and BLOGs) agree there is some ROI in the flexibility that IaaS brings. In the article you reference, I think he was calling that "cloud bursting". I'm assuming he is talking about the ability to quickly scale up for high demand and then return all of the unneeded capacity when it's no longer needed, saving the fixed cost investment to scale "on-prem".

      What I'd love to see are costs associated with growth that busts the data-center walls when you need growth. There is a big step function in cost and a hell of a lot of time to do enlarge a "on-prem" implementation. If you find you no longer need the bigger data center, then you are kind of stuck with it.

      I know this CIO that had to spend Super Bowl Sunday moving his data center because of a real-estate decision made by the bosses. Would have been nice to be using IaaS in that case :)

  2. While IaaS might have allowed that CIO to enjoy the Super Bowl instead of moving the DC, enjoyment of the Super Bowl was probably not a KPI that drives his business. I'm sure he is dealing with the same tough selling points surrounding the Cloud that every other CIO faces: relative trust in security, SLA performance, concern over vendor lock-in, etc. Flexibility/growth is a very compelling story for IaaS, but that attribute may not always be tops on the list for every company.

    I'm enjoying the series. Please continue your exploration of "Why Cloud?" by looking at the whole range of concerns and mitigating factors.