And the Cloud Goes Poof!

Not really, because there are an awful lot of people committed to making sure that you do your computing remotely, but halfway reading this PC Magazine article (via Yahoo! News), I started thinking, “hey, this sounds like Dvorak.”  No surprise, it was written by him.  He’s been very, and loudly, critical of the Cloud over the years, and the Mass Sidekick Annihilation of 2009 is the perfect opportunity for him to take another swipe at it.  As I am on the record as being pretty anti-Cloud, I of course like the article.  But this post isn’t about his article (which even takes EULAs to task), but about the damned Cloud.

I think the lesson to take from the Sidekick meltdown is that this may be the first time, on a large scale, that a failing Cloud has actually caused serious loss of data.  Whereas previous failures of the Cloud yielded irritation (such as GMail outages and Amazon down-time) this is a failure that actually hurts people where it matters: for a business person using a Sidekick, contacts are money.

Some people, it must be noted, are saying that the Sidekick data loss should not be considered a failure of the Cloud.   Tony Bradley, for example, says:

When Google News or Gmail experience outages the media ([he] included) cries that the cloud has failed. When Danger loses Sidekick data synced online from a million users the media declares it another failure of the cloud.

But, Google is not the cloud. Claiming that a failure by Google is a failure of cloud computing in general is like claiming that New Coke was a failure of all soft drinks. Declaring the loss of Sidekick data a failure of the cloud is like claiming that the failure of the Yugo was a failure of automobiles in general.

He rightly points out that the problem appears to have been improper server maintenance and the failure to adequately back up the remotely-saved data.  And it’s true that–for right now–things like Google Apps and remote synching of phone data are somewhat like New Coke and Yugos: the tech is not yet in wide enough adoption for one to be able to say that New Coke is the only soft drink choice (not that I ever had a problem with New Coke, but I guess I’m a little cola ambivalent when it comes down to it) or that Yugos are the only car choice.  But the impetus behind Cloud computing and SaaS is that it is intended to supplant rather than supplement local storage and computing. (See, e.g., this article in Australia’s The Age.)

The Sidekick has about 1,000,000 users, according to the articles discussing the outage, and apparently there isn’t a convenient way to store the data locally.  Contact info and calendar entries can’t possibly take all that much space on a server, so it should be relatively easy to maintain server health for such a small amount of data.  Think about the problems inherent in trying to maintain server health when hundreds of millions of people are storing much larger amounts of data in a remote location.  And that’s assuming a benign ecosystem, one where there aren’t constant attacks on the storage facilities.  As I noted the other day, however, that isn’t the reality as the Cloud is a prime target for malevolence.

Slippery slope?  Yup, in a way, but by the same token it’s a real problem and I still can’t get behind the idea of giving up local computing.

1 Comment

  1. It is simply about Reputational risk. Reputational risk (damage to an organization through loss of its reputation or standing), can arise as a consequence of operational failures. Every company understands reputational risk, particularly businesses who regard their brand as one of their most critical assets. Google is one of them. They have a reputation to maintain. But I suppose Microsoft Danger didn’t care…

    Ability to export data and store it locally is extremely important. Cloud services need to provide that ability. From what I seen (and experimented with), Google provides excellent set of APIs to access the data stored in Google’s Cloud. And Google is always working on to improve the APIs. Google usually first adds functions to the API, and then introduces them in the UI. Compare this to other software vendors, who usually introduce the new functions in the UI and then at a later time provide API access to those functions – if it all.

    I use both Google Docs and Windows Live Workspace to store my personal / school related stuff. I use both of these because they both have their benefits. Windows Live Workspace provides complete integration with Office 2007, whereas Google Docs provide editing capabilities in a Web browser. Recently I have been thinking of writing an application that will synchronize the content of both of these repositories. Google provides APIs that make this task easy from Google’s side, but there are no Windows Live Workspace APIs, so I have to devise a workaround to get documents into the Windows Live Workspace.

    oh btw, also check out Data Liberation Front

Comments are closed.