Posted by: Arun Gupta
Add new tag, cloud and the CIO, cloud models, cloud outages, cloud service providers, SLA
The beginning of the monsoon season in Mumbai inspired me to push the boundaries again in quest of the silver lining in the cloud. Recent events around outages and security across multiple global cloud pioneers poses doubts on the movement of even non-mission critical applications outside of the corporate data centers. We are not just talking about infrastructure or platform as a service, but everything that is the manifestation of the public cloud.
Evolution of cloud
Over the last couple of years, every offering saw two shifts: first it had to have a cloud flavor and second around social networking (that is another story). Some termed this new euphoria as bubble 2.0 tinted by valuations achieved in recent IPOs. So everyone justified how this time it is different and why it is sustainable. Many large and small enterprises found efficiencies, at least short term, in shifting field functions like sales and service and collaboration on the move to the cloud.
Leaving aside the debate between public, hybrid, and private clouds, the real issue is about the promise of the cloud, irrespective of the vendor, the type of cloud offering, or the engagement model. The big benefit that every type of cloud offered was savings, real quantifiable savings, or better, total cost of ownership. CFOs would agree that TCO is always a good measure for any financial model if all other dimensions remain unchanged.
Cloud service providers’ financial models are contingent on multiple customers adopting their base-solutions which give them the efficiency of scale and repeatability. As the number increases beyond a threshold, they start making money. Non-concurrency improves yields, but prices remain the same for customers. So the financial models attempted to capture some efficiency-based gains, making them look attractive to the prospects.
A few unanswered questions
Most discussions got off to a good start with worksheets providing easy decisions. The newness of the paradigm left some questions unanswered, but during the slowdown, these were brushed aside. Some of these were:
a) What happens if the SLA is not met?
b) Is my data as secure as it is in my current state?
c) Can I move off to another cloud if I don’t like something? How easy is the transition going to be?
d) As I upgrade the internal systems, how do I ensure that their integration with the external systems does not break?
e) What recourse do I have if the cloud service provider goes bust?
I will stop here, the list is a bit longer, but you get the point.
Business impact due to recent outages and security breaches for some of the smaller customers was significant. Some of them just had to wait and watch with no option. A few had spread the risk across, and thus, the impact was limited. The big enterprises shrugged and moved on. How does one balance the adverse business impact against the cost savings? To me this is a bad compromise as everything is subservient to business interest.