Risks in Cloud Computing

On May 14th US timing, Google experienced a wide-spread outage lasting several hours that affected many of their services like Youtube, Gmail, Google Analytics, Google Maps and Google Docs. The outage affected many countries like the US, Europe, China and Australia. This comes right after the outage for Microsoft Azure cloud.

Considering the scale of Google's services, it's commendable that they were able to bring up all the services in a few hours (faster than Microsoft 22 hours outage). However if you look at the other perspective, even a huge company like Google could suffer an outage, who else will not?

Recently I've seen many articles written on cloud computing, both abroad and local. First of all, what's the foremost reason why you want to consider cloud computing.

It all comes down to if you want to outsource your IT operations. BUT... IT often drives productivity in an organisation, if implemented properly. Nowadays, what differentiates the competition is your response to the happenings in the business world. And very often, this response is inter-related with your IT infrastructure, and how you use IT.

Just imagine if you put all your company's data, email, etc in the "generic cloud". What happens if your cloud is down several hours like Google? What will happen to your business operations? What happens if let's say you're running a mission critical system? Using an example of a healthcare system, can you imagine the implications on such an outage? Of course you can say that you're protected by the SLA, 99.5% uptime, etc etc. However, the outage can cause irreversible damage to your company's reputation.

Some will argue that you will also face the same problem maintaining your own IT infrastructure. However, you have to take into account the complexity. Your cloud computing vendor might have 1000s of customers. Can you imagine how complex will their infrastructure architecture be? Compared to your ONE company's infrastructure? Coming from the technical standpoint, I personally believe the cloud infrastructure is much more complicated to bring up once it goes down.

I've also seen articles on hybrid clouds, which I also do not agree. Other than outsourcing your IT operations to the cloud, the other advantage of the cloud is the utility model. You're only paying what you're using. It makes absolutely no sense for example for the vendors to provide only the application hosting, and not the actual database hosting. They will never commit to any SLA because access to the application and access to the database is interconnected. It HAS to come as a package.

That brings about my next point. Security... Coming from the perspective of economies of scale and cost, I seriously do not see the cloud computing vendor deploying separate databases for each customer. The most likely scenario would be a huge database with many schema users. It just takes ONE unsecured schema user and a determined hacker, and your data can be compromised. So what if there is high availability, backup, uptime, etc?

Previously back in March, I've written a post related to this topic. There's still no change in my position. I still strongly believe that the generic cloud can only be used for public services. For example, your company's website, forums, etc. If you want to put companies' confidential data like CRM, ERP, etc, I still think a "private cloud" will be the way to go, combine with the use of virtualisation.

There's also one critical fundamental question you need to answer if you wish to go into cloud computing. What happens if the cloud computing company goes out of business?

Something for you to think about...