Wednesday, July 9, 2008

5 Reasons to avoid Cloud Computing

I will follow this post with a "5 Reasons to embrace Cloud Computing". I am actually very pro-cloud computing. I think it is the next big thing.

But, and there is always a but, there are reasons to avoid it right now.

  1. It's immature - we don't really know who the long term players will be. We can guess on MS, IBM, Google, Amazon. What about Mosso, GoGrid, Elastra, Rightscale and any of a dozen others. The market needs to mature and play out for a while.
  2. Security - is the cloud secure? I think it is actually. The problem is that I have yet to see any of the infrastructure providers offer any accounting or auditablity of that security. That is a barrier to compliance.
  3. No disconnected computing - what happens when the construction company fixing the road cuts your cable connection? Ok, probably the same thing as if you used a data center. But what happens if your cloud is based in the middle east, or africa? What happens when a ship's anchor cuts a cable? It could be. Most providers won't tell you where your data and applicatiosn are sitting.
  4. How much does it cost? We know how much it costs today but how much will it cost tomorrow? No one is offering long term contracts for infrastructure because they don't know what it will really cost. Third party management and tools vendors are offering subscription pricing and that gives you some protection but if the cpu per hour cost at your provider goes up, what will you do with those subscriptions?
  5. SLAs! This is my particular peeve of the moment. Amazon offers a future discount for not meeting their SLA. That's not adequate for most businesses. However, coughing up cash for a major outage could put a provider of business. I wonder if cloud insurance might be on the way in the future?

Those are my 5 reasons to avoid the cloud for now.


1 comment:

Mario Meir-Huber said...

Hi Lewis,

just came across your post as i am doing some research in that field; they match my facts i used in my book, but i have some more for you:

- screening. what happens when governments reserve the right to access data? what happens if the us law requires european company data to be analysed? industrial spying might become easier.
- bankruptcy. we can't really predict what companies will survive. imagine microsoft will fail, as they can't sell windows or office any more ... i think it is very unlikely but we never expected the banks to fail ... but they did. if a company fails, what happens to my data or apps?
- vendor lockin. if i want to switch, it might not be easy ... every plattform uses different apis and so on .. imagine you have 100 TB of data on a plattform and you want to switch to another one ... that's not gonna be fun ;)
- costs. iaas and paas costs are hard to predict/estimate. i would say it is even up to the app architecture how much someone has to pay. with all the storage operations and costs for traffic it is hard to say how much a plattform will cost.