18.5.11

What could possibly go wrong (with Cloud Computing)?

We all know one... The eternal optimist who's absolutely determined that everthing is fine, that you worry too much, that you should just unknot your knickers and get on with life.

He constantly reminds you that things have a way of working out in the end, and that 80% of what we worry about never actually happens.

Actually, the world needs more people like this, just not in the role of CIO.

As IT leaders, we have a complex role in that we are constantly balancing passion, hope and building vision with anticipating anything and everything that might go wrong.

If I look at my job description, this is embedded right in between developing strategic direction and leading my team. The area I'm referring to is Business Continuance and Disaster Recovery. When things go wrong, IT is usually one of the first responders. Who gets to stay up all night so that work and production are not interrupted when things fail?

This is of course no secret, and IT folk take this responsibility very, very serious.

It's rarely the big things that we deal with. I mean, when was the last time you had a building blow up? You may have had a network switch or server fail, or a fire or flood. In my last few months my team has dealt with a major electrical short that took our whole campus offline for a few hours. But all of it was easily dealt with by our continuance plans. Pretty much everything we have had to deal with has been under our control. We had access to our data centre, our switch closets, our facilities, our backups, etc.

But cloud computing introduces some significant changes to our Business Continuance strategy.

We are told that cloud computing provides some significant benefits:
- Our infrastructure can be scalable as we need it. We only need to pay for what we need.
- "Patch Tuesday" is now a thing of the past. The servers are kept up to date by the cloud provider.
- Our data is safe, backed up and available anywhere there is an Internet connection.
- and so on... but this isn't a post about the benefits of cloud computing.

Does cloud computing really make my BC/DR strategy easier? I mean, what could possibly go wrong?

What happens when the cloud goes away? Like the recent events with Amazon and Blogger?

In Amazon's case, several companies hosted on Amazon's servers were completely offline for more than a day. Blogger not only went offline, but lost 30 hours of postings.

Even offline storage from a large company can go away. Cisco has recently sent an email to its Flip camera users that online storage of Flip photos would only be available for 30 days. If users didn't copy the photos to a local drive, they would be deleted.

If you are like many CIO's, this would be enough reason to not put things into the cloud. How do you recover from these types of problems if you have no control over the environment? How can I recover from a server problem if I don`t have access to the Data Centre?

If you were planning on putting your whole computing infrastructure into the cloud you may have a valid concern.

Prudent planning for the cloud will encompass your business continuance strategy. At the recent Midsize Enterprise Summit I attended, Gartner`s analyst stated that cloud computing may take up to 10 years to fully realize the benefits, and the best strategy to approach this is to start with elements of your business that are not absolutely mission critical.

If you do commit mission critical applications to the cloud, then planning for potential issues is critical. When Amazon lost their data centre, Netflix stayed operational since it had spanned its application across more than one data centre. For those companies that only had their application hosted at the affected data centre, they ended up unavailable to their customers.

It's impossible to anticipate everything. Dwight D. Eisenhower said "I have found that in the midst of battle, plans are useless, but the planning was indispensible."

So, if you are moving to the cloud, my recommendation is that you abandon your inner optimist (for a moment), channel your inner Eeyore and anticipate the potential issues that may impact your ability to quickly recover from disaster. I`m not saying it will be easy, but it is very doable.

Once you have that documented, then resurrect your optimist (think Tigger) and start selling your organization on the benefits of the cloud.


This post was one of my entries in Computerworld Canada's 2011 Blog Idol contest.

1 comment: