Gartner Predicts the end of the PC by 2014, and this time we mean it.

On Wired Cloudline, Mike Barton covered Gartner’s prediction that the “Personal Cloud” will replace the PC by 2014. I think we should call it the “Networked Personal Cloud” or “NetPC.” which has been predicted to replace the home computer since 1996. Gartner lists “five megatrends” that point to a mass migration to the cloud, while missing the real big picture:

Megatrend #6: Consumers are burned over and over again by loss or violation of their personal content in cloud providers and really start to understand that if the data isn’t on your desktop, you don’t own it.

While there may be market pressures encouraging the Cloud/dumb terminal architecture, corporate short-sightedness may seriously dissuade customers from moving. These are some of the recent writeups on fallout from storing personal information in the cloud:

Jeff Vance wrote about the Top 10 Reasons Cloud Computing Deployments Fail (Datamation, July 2010). A vital point is #2, which details a web company that went down when Amazon EC2 went down. He points out that the cloud is a single point of failure. While there are things you can’t change about this (“Well we’ll just install another internet.”), you can mitigate it. For a business, this might mean not hosting your public and support websites with the same host you run your business from. As a consumer, this means “don’t put your only copy in the cloud.” And if you’re not putting your only copy in the cloud, then you’re going to need a PC to store it on.

The Cloud is a Single Point of Failure

In addition, having a dumb terminal and everything in the cloud means no internet = no computer. I remember that there were internet outages during the 2010 blizzards, and a number of friends who were heavy “cloud” users couldn’t do anything. No GMail, no Google docs, no online code repositories, no streaming video… Nothing. Meanwhile, while I couldn’t send or receive, I could work on the email I had in Outlook, all my documents worked fine in Word and Excel, and the digital media files I had in my home server played.

We also don’t know the infrastructure effects that face us. While we’re putting more and more “stuff” online, there are signs that “the internet” is struggling to keep up. Comcast capped data transfers, and mobile phone providers are throttling “unlimited data” plans. So while Google, Netflix, and others want us to consume more online information, the companies that provide the pipes to this information are making them smaller.

Personally, for a long time I have thought that too many folks were “all or nothing” and missed the best solution around. Microsoft tried to call it “Software + Services,” but they never sold it right. I call it

The Outlook/Exchange model

Anyone who’s worked with Outlook in recent years knows how easy it is to work with. You have software installed on your PC that you use to read and write mail (and schedule appointments, and your address book). When you get an email, it’s downloaded and available on your computer. You can read, reply, search, whatever. If you’re not connected to the internet, you can read and reply to mail – it’s just stored until you’re connected, at which point it’s sent without you having to do anything. Meanwhile, interacting with Outlook is clean and snappy, even if all you have is a dialup connection.

However, consider the email you have locally just a copy – the “data” is “in the cloud” – on Exchange. You have the local copy, and can even back it up, save it, move it around, etc. But when you get a new computer and install Outlook, just enter your email address and password – it connects to Exchange and downloads all your mail. Come back in a little while and everything is just as you remember.

In addition, there’s Outlook Web Access, or OWA – if you’re on a public terminal, you can open a web browser and go to your company’s Exchange server, where you’ll be shown your mail in a mailbox that’s very close to Outlook.

Not Cloud – Cloud/PC

This is the right model for applications in the future – a cloud backend, but an installable front end that reduces minute-to-minute dependence on the Cloud. It reduces bandwidth requirements and loading on the infrastructure, provides a better user experience, and is even easier to program for. You can still offer a web application for folks who want it, but don’t make them dependent on it.

  • Jeannette Westlake

    This is a great paradigm that is simple to implement, rational, and avoids all of the obvious pitfalls of cloud-based computing.

    It’s a really good idea…

    Hence, I seriously doubt that it will become the dominant model.

    For some reason, “What happens when the power goes out?” never seems to occur to the people ultimately responsible for critical infrastructure. Just look at Fukishima.