alanwilliamson

Engadget claims "biggest disasters in the history of cloud computing"

Cloud Computing eh? Love it or hate it, chances are, you probably don't know what it is. Yet you feel comfortable branding anything that runs inside a browser as "cloud computing". Why wouldn't you? You got bored of "Web2.0".

It is trendy, it is the latest buzzword to be doing the rounds, and like everything else in the computing world, if you can just get that edge on your peers and show how "in tune" you are to the latest pulse people will look upon you more favorable, and hopefully untold riches and offers of people wanting to sleep with you will rain down upon you.

Engadget, under the stewardship of Chris Ziegler, published their reporting on T-Mobile's unfortunate problem of where it looks like they may have lost all their users Sidekick data.

Not content with merely reporting the facts, Ziegler had to juice it up a bit and include in his opening sentence no less, a complete side swipe at cloud computing, laying the blame at its door. Excuse me?

What on earth has this got to do with cloud computing?

I'll tell you, just in case you are wondering, absolutely nothing. Does running out of petrol in your car, mark "disaster" for the automobile industry? Does missing a flight, mark "disaster" for the airline industry? Does failure to connect to a server, mark "disaster" for the Internet?

Havana out of fuel

Of course not and neither does T-Mobile failure to backup a server have any impact on cloud computing let alone mark it as its darkest day. No where on T-Mobile's website do they claim they are using "cloud computing".

What T-Mobile was providing here was a service. No different to your bank providing a service for your money, or the postal service for your mail. You entrust organizations to perform a duty on your behalf, and if they fail to deliver, then yes, you have to lay the blame at their door. Yet if you do rely 100% on that service, then you must take some responsibility in that contract, always looking to protect yourself against that worse-case scenario.

Cloud Computing is not new. We've been living with "cloud computing" (for some peoples definitions) ever since the first http server ever started listening for requests on Port80.

But cloud computing is more than merely providing a service. It is about the auto-provisioning of services when you need to scale up (or down) or to replicate due to failure. It is about providing a seamless service, and utilising the underlying platform to do the things that were historically difficult in traditional data centres that are tied to physical machines. Cloud computing free's us from the physical chains and allows us to build-in redundancy and failure from the start.

The industry is still trying to formulate a definition for "cloud computing", because we are always running up against this ignorance to label everything inside a browser as the wonders of the cloud.

Engadget could have easily have rewritten their attention grabbing sentence:

  • "biggest disasters in the history of the internet"
  • "biggest disasters in the history of the web"
  • "biggest disasters in the history of the mobile phone"
  • "biggest disasters in the history of the network"
each of them as relevant as the one they chose to use. They are not the only ones at fault here, they just happen to be one of the first ones out of the gate with the cloud angle, and others have followed suit.

The next time Google has an outage, HotMail inaccessible, or your favourite website goes offline, then put it down to what it is, a "breakdown". We seem to want to point the finger of blame at something bigger than the company at fault.

These the joys and warts of utilising a service and should come with a warning: "the value of your service can go down as well as up".

Comments

@franktronic: it really depends on your definition of what cloud computing is, and if you take it to be one that is anything in a server, then yes, it is related.

That is the problem facing cloud computing at the moment, trying to define what the actual definition is. Two of the most outspoken advocates of cloud computing (Sam Johnston and Reuven Cohen) can't actually agree on it.

Sam: http://samj.net/2009/10/if-its-dangerous-its-not-cloud.html

Reuven http://www.elasticvapor.com/2009/10/cloud-computing-is-dangerous.html

left by Alan Williamson — Monday, 12 October 2009 9:07 PM — web site

"What on earth has this got to do with cloud computing? I'll tell you, just in case you are wondering, absolutely nothing."

*Nothing* at all? Really? Maybe Chris's hyperbolic remarks were a bit of a stretch but to say that this has "absolutely nothing" to do with cloud computing seems even further from the truth. Simply storing data in a nebulous, internet-connected server may not be cloud-based SaaS but it sure feels *something* like cloud computing to me.

left by franktronic — Monday, 12 October 2009 8:54 PM — web site

T-Mobile can barely keep any system running, and that's not fault of the system.

left by Kirk — Monday, 12 October 2009 8:06 PM

@Randall, such "service" failures like these are not new. The service industry is littered with tales of lost photos, files, emails. Any company that takes on the responsibility of being the primary holder of a users data has a duty of care to take extra pre-caution and to always plan for the worse-case.

This is a responsibility that has existed ever since we started using service orientated applications. In other words, this pre-dates cloud computing by a long reach.

My argument is not to lay the blame at the cloud computing door step, but instead look at it as a problem of any service we rely on. Users have gotten too complacent with their trust and faith in 3rd parties. Steps should always be taken to have a backup to the backup.

Yes, in this particular instance T-Mobile have a lot of questions to answer and it will make people more wary of using T-Mobile/Sidekick in the future. But will this make people more wary about other service providers?

I doubt it, but I hope it will make them think-twice about the data they are loading onto that service and ask themselves if they could happily survive if that data was suddenly unavailable.

Now, if you really want to bring cloud computing into this argument, then one can present the case that a proper cloud architecture may have been able to reduce the likely hood of this massive data loss, as Sam Johnston makes the case for here http://samj.net/2009/10/how-open-cloud-could-have-saved.html

left by Alan Williamson — Monday, 12 October 2009 4:01 PM — web site

I'm more inclined to believe that this is one of the biggest large scale failures of infrastructure to date. You're saying that a sidekick customer who was burned isn't going to think twice about trying cloud computing / storage / whatever you want to call it again?

left by Randall — Monday, 12 October 2009 3:53 PM

Leave Comment

please note, all comments will be moderated for spam and abuse before being publicly posted.


 

Recent Cloud posts

Recent JAVA posts

Latest CFML posts


 
Site Links
Recommended Sites/Blogs