Shhhh Stuff Happens…In the Cloud

November 13, 2009 by Brian Wolff

By now everyone has heard about the T-Mobile Sidekick crash in one of Danger’s (a Microsoft subsidiary) cloud computing environments.  Just like every time Gmail goes down, there is a wide and broad brush of FUD painted across the cloud computing canvas. 

Stephen Foskett recently wrote a blog post that talks about the shocking “bubblegum and baling wire” approach that some companies take when delivering cloud computing storage.  I couldn’t agree with him more.  There are some hokey things going on in the world of cloud storage, otherwise these crashes wouldn’t happen and James Urquart wouldn’t be questioning the need for malpractice safeguards in cloud computing.

As the title of this blog post implies, this type of data loss that occurred at Microsoft is more common than many would admit.  The difference is that most times it’s not in a public cloud and it doesn’t affect thousands of people.  Why does it happen?  It happens because moving parts break and because human beings are involved.  Now, just because there are breaking parts and humans, that doesn’t mean there must be a disaster or data loss.  The question is really about whether your business is willing to pay for what it takes to insure that the data is protected or in the case of Microsoft Danger, take the extra step of providing a path to copying the data to another instance.  Data loss can be prevented; it’s just a matter of who’s willing to pay for it.  It would be wild speculation on my part to say what actually happened in the MS Danger data center – but what I can say with much confidence that an economic decision was at the heart of the loss. 

How do I know that?  Because at BlueLock we take data loss very seriously, so we have engineered redundancy into our cloud architecture and that kind of protection comes with a cost.  Some prospects choose to pay for it (we call them clients) and some don’t (we call them “lost deals”).  Go here to learn more about how we back our data up.   More often than not, the prospects that choose not to pay the cost do so because the data they are trying to protect is not worth the cost that I’m proposing to keep it safe.  Of course the converse is also true, our clients that accept the cost, do so because when they see all of  the engineering and services bundled into the price, the cost is in balance with the risk of the data loss.   

The bottom line…

Just as I discussed in a previous post about the TCO of an Infrastructure, there are many costs associated with protecting data and insuring that failing hardware or humans don’t take you down.  The bottom line is really about the risk of loss and a careful cost/benefit analysis of paying to protect the data….or not.  If you’d like help completing that analysis, send me a note or give me a call.  That’s what we do every day for companies just like yours.

Reference post by Stephen Fosket