Hey Chicken Little, the Cloud Is Not Falling

Written by Alex Williams / October 12, 2009 12:39 PM / 14 Comments

We sure are hearing the chickens running around in a panic about the dangers of cloud computing following the massive data loss involving T-Mobile Sidekick customers. And as usual, the cacophony sounds more like a bunch of pundits ruminating about the great dangers that may be ahead instead of the reality at hand.

The problem is, most of them are making zero distinction about what constitutes a cloud computing service. The Sidekick disaster was not the result of a cloud disaster. It was a centralized data center that had poor oversight.

DevCentral clears things up with its distinction between cloud services and applications:

A "cloud service" is used by IT, by developers, by the technical community at large. What consumers access is an application, and nothing more. They aren't the user of the cloud service, they are the consumer of an application deployed in a cloud environment. Google Docs is an application. Gmail is an application. Twitter is an application. None of these are "cloud" services, even when using APIs designed to integrate them with other applications; they are still, always and forever, applications.

We do not question the severity of what happened to Sidekick customers. It looks like about 1 million people are affected. They lost it all. Pictures, calendars and a whole host of information.

These customers had no choice about what happened. They relied on T-Mobile. And T-Mobile relied on Microsoft/Danger for storing the data. This was not a cloud catastrophe. Developments continue to unfold: Hitachi Datasystems is now being fingered as the source of the problem.

But since we are on the topic, there are some basic lessons to learn in working with a cloud service provider. This is not a complete list. Feel free to add your own pieces of advice.

Lesson #1

If you are storing your data in the cloud for your customers to access, you'd better know if the company you have hired is actually the one managing your data. If your vendor is outsourcing your data to another provider, it could be a recipe for trouble.

Lesson #2

Know who you are working with and make sure there will be no surprises. Software-as-a-Service (SaaS) providers that don't keep customers posted about changes or upgrades can be real trouble-makers. All kinds of mix-ups can occur. A SaaS vendor recently pulled this one on its users. Customers had no idea about the upgrade. They had no control.

Lesson #3

Make sure your provider has safety valves in place. How is the data backed up? Let's say, again, that the SaaS provider does an upgrade, but there's a nasty bug fouling things up. If the cloud configuration has a safety valve in place, then the customer can mitigate the issue pretty easily.

Lesson #4

Don't use just one cloud service provider. Security experts make the point that you don't put all your eggs in one basket. Look at multiple cloud service providers so that if there are issues, damage is limited.

Perhaps, overall, the greatest lesson out of the Sidekick disaster has nothing to do with the cloud at all but more about the applications that people use in the enterprise. Facebook? Twitter? Those are applications that may be more troublesome than cloud computing services because of their vulnerability to attack and lack of control over the data.

Comments

Subscribe to comments for this post OR Subscribe to comments for all ReadWriteEnterprise posts

I politely disagree with your rather strict interpretation of "cloud computing". Cloud computing is not just about applications, but storage and services as well. Danger's infrastructure was a storage cloud by any reasonable interpretation.

This is clearly a "Cloud" failure... and a perfect example of what appears to be an infrastructure design gone bad.

Despite news reports saying crazy things like "All data only on one server" and "no back-ups", I believe those things are going to end up being untrue. What will be true is a backup configuration and strategy to closely entwined with the master data.

Cloud computing has it place and provides tremendous value, but clearly dangers exist (no pun intended), and any company or project must do its due diligence before embarking on that path.

Posted by: chanur | October 12, 2009 1:23 PM

It seems ridiculous to me to imply that the fault here lies with the customers who failed to do their due diligence in selecting a provider, or to marginalize the seriousness of the failure due to a technicality based on a clinical definition of cloud computing. The idea of "cloud services" is only a small part of the bigger picture of cloud computing as a whole.

The fact is these customers had trusted MICROSOFT with their data, and as far as I know weren't given a choice of options for providers. Regardless of who bears the blame for the hardware failure, there is NO excuse for not having redundant systems and working backups.

I would venture to guess that facebook and twitter both have their critical user data safely tucked away.

Posted by: Mark Carraway | October 12, 2009 1:30 PM

@chanur

From Om:

"Microsoft was upgrading their SAN (Storage Area Network aka the thing that stores all your data) and had hired Hitachi to come in and do it for them. Typically in an upgrade like this, you are expected to make backups of your SAN before the upgrade happens. Microsoft failed to make these backups for some reason. We’re not sure if it was because of the amount of data that would be required, if they didn’t have time to do it, or if they simply forgot. Regardless of why, Microsoft should know better. So Hitachi worked on upgrading the SAN and something went wrong, resulting in its destruction. Currently the plan is to try to get the devices that still have personal data on them to sync back to the servers and at least keep the data that users have on their device saved. We’ve heard this from what appears to be several sources and it seems to hold weight. Needless to say it all boils down to one thing: Microsoft did not have a working backup."

http://gigaom.com/2009/10/11/microsoft-mobiles-worst-week-ever/

@Mark Carraway

"there is NO excuse for not having redundant systems and working backups. "

+1

@ all Microsoft apologists

This is not a failure of cloud services, not a failure of the users, most certainly not a failure of the LAMP stack...

It the status quo - Microsoft and its total contempt for the user.

Let's go ahead and call it now:

MICROSOFT’S AZURE IS DOOMED BASED ON SIDEKICK DATA LOSS

Posted by: Todd | October 12, 2009 1:36 PM

What's making this event so interesting is not the "falling cloud"...it's all about the "Microsoft failure".

If want to run your services on Azure next month, it's better to make the backup by yourself.

Posted by: GRO | October 12, 2009 2:10 PM

Here's three great secrets to success in life, school and business...back up, back up, back up!

RonD
http://www.start-a-business-faq.com

Posted by: Ron Derven | October 12, 2009 7:54 PM

I think one good way for customers to gauge a SaaS company is by checking its customer service out. If it is quick in responding and fixing the bugs then it probably has taken care of security, backup etc..

Posted by: Priyanka D | October 12, 2009 11:35 PM

I never liked the Sidekick anyway, I was always a Blackberry owner...very faithful...like the rest of the world.

Posted by: edwin sanchez | October 13, 2009 6:32 AM

Thanks for this post

Posted by: computer repair Toronto | October 13, 2009 9:55 AM

I think that they did a great job at first but now they really need to get it together.

Posted by: Chris Brown | October 13, 2009 11:21 AM

Thanks for taking the time to write this and help put an end to all the hysteria around this. Outages in the cloud do not happen often but they are highly publicized. Just like the hysteria over "swine flu" will soon fade from memory so too will this.

Posted by: islandinthenet.com | October 13, 2009 11:26 AM

I agree with Alex's post. Facebook and Twitter may be more troublesome than cloud computing services because of their vulnerability to attack and lack of control over the data.

Posted by: Fango Ono | October 14, 2009 8:38 PM

The red flag for people considering the cloud here is that while the cloud is independent of single pieces of hardware, it's still served from hardware.

When hardware requires maintenance, outages/downtime can occur. Cloud services are generally designed to continue operations in the midst of individual hardware outages, but with such a new technology, it's not uncommon to find single points of failure that affect service (even if the single point of failure happens to be a single SAN that hosts the entire cloud).

As small- and medium-sized businesses look to clear out their server closets and outsource their IT, the first network hiccup or bout with downtime will cause a similar rift in their confidence in the outsourced IT model. The benefits of the decision aren't necessarily noteworthy, but the black eyes are glaringly obvious. If Sidekick users lost their phones and could repopulate their replacement device with information from the cloud, I'm sure they were ecstatic ... but they probably didn't make T-Mobile or Danger or Sidekick a trending topic on the highest trafficked tech outlets on the Web as a result.

It'll be interesting to see how much of the data could actually be recovered (and how that plays into the consumer perception in the long run): http://forums.t-mobile.com/tmbl/?category.id=Sidekick

Posted by: The Planet - Kevin | October 15, 2009 8:52 AM

Interesting lessons...
Vendor's domain expertise and rich exposure to real time customer experience can be another critical aspect while selecting an application provider in cloud

Posted by: Sreekumar J | November 2, 2009 10:54 PM

It's pretty sad that my small company's network and data is more redundant than a Microsoft network. We have live, on-site backups using VMWare, then geographical live backups in an offisite datacenter. If we have a fire or tornado destroy our building, we can be back up and running in less than an hour from a free coffe shop WiFi connection if we had to!

Posted by: tastaturen | January 7, 2010 2:37 AM