What is Cloud Computing?

March 18, 2010

Question: We’re running a news service with volunteers here at PullNews. We want to know if you think it’s valuable as a web page (a la TechCrunch for semantic web) or if you would be just as happy getting our tweets on Twitter and forget the HTML version here on the site. We’re asking for opinions. Please contact one of us or just leave a comment below. Thank you.

I want to talk today about cloud computing, something many people don’t understand. They confuse it with online storage. They confuse it with pretty much everything. There are now as many definitions of cloud computing as there are consultants.

First, Larry Ellison talks about cloud computing, saying it’s gibberish. It’s certainly true that there is a lot of gibberish about cloud computing out there. And, as usual, Steve Ballmer is the last person to figure out what cloud computing is. One nice guy explains very adorably how cloud computing is cloud hosting, which i think is a good start but not the whole story. A lot of people say it’s a collection of expandable online resources that enable Software As A Service. That’s closer, but not what I’m looking for.Take a look at Cloudo. It’s kind of a typical startup in this space. It provides storage, so you can “collaborate” with colleagues by sharing data files. But the apps don’t really run online. It’s just storage. It’s a far cry from being able to collaborate as I describe in my “I have a dream” blog post.

Now look at EyeOS, a company that IBM has chosen to provide cloud computing architecture on their new servers. They EyeOS guys have written something important on their blog I want to share with you:

The main difference between EyeOs and the majority of cloud computing services is that they ask for your data and they give you the service in exchange for it. Because EyeOs is free software, you can decide where you keep your server and, more importantly, your data. We do not ask you to send us your data and you can install EyeOs in an environment that is completely under your control.

Did you see that? You can store your data anywhere. Rather than mimicking the old server/application/database methodology, you’ll be able to use EyeOS with distributed data, like linked data. That, to me, is the key to the future of cloud computing – that the data is in the cloud, and the apps can call the data by name, or that the data can call the apps by name. The latter approach enables pull. When the data doesn’t move, the apps can live anywhere and can be called to the task by their names. As I keep pointing out, we often lose our data, but we never lose our apps. It should be the other way around. Apps are easy to regenerate, because they always have back-ups. Data, which flies everywhere and we deal with dozens of times per day, is much easier to lose. That’s pretty promising. Perhaps it will help evolve into the operating system necessary for the design and collaboration infrastructure I talk about in my “I have a dream” post.

Some people at Google are developing an operating system that goes way beyond their fast browser. It’s called the Google OS, and you can learn all about it at Chromium.org. It’s probably the closest thing we have to a true cloud-computing platform today.

Spend some time on theCloud Computing page at Wikipedia, and you’ll get an eyeful of computer jargon that basically boils down to flexible hosting and processing power that enables SAAS.

So to make things more confusing, here’s my definition of cloud computing, for the record:

Cloud computing is both a scalable online application platform and a scalable online data platform. Applications can call data by name, which means that the data can be located anywhere, and it could be distributed and linked on hundreds of servers. More important, the data can call applications, so data can stay where it is and the applications can come and go. To enable the semantic web, cloud computing will support the creation of ad-hoc applications built of several smaller tasklets to tackle a given problem without writing custom code.

That, to me, is the real goal of cloud computing – that our software no longer comes in huge applications but rather in tiny tasklets that specialize in doing one thing only. The tasklets can live online and show up as needed. The data can be assembled from the cloud as needed. To solve a particular problem, everything comes together as needed, then goes back where it came from. See my “I have a dream” post to get more of a feel for what I mean. (it’s just below the Twitter post, below).

One guy who is ahead of the crowd is Paul Miller, whose blog is called Cloud of Data – sounds like he has his priorities straight. He interviewed me about the power of pull a few days ago. Our conversation is on ZDNet this morning.

Related articles by Zemanta

• Open Cloud Sync and Standards Matter More Now Than Ever (ostatic.com)

2 Responses to “What is Cloud Computing?”

Follow comments with RSS 2.0

1. 3/19/2010 Update « CloudRoad | March 18, 2010 at 10:06 pm

   [...] The Semantic Web & THE POWER OF PULL Blog Archive Cloud Computing [...]

2. 

   JZA | March 22, 2010 at 2:31 am

   You should check out Eben Moglen on the Freedom of the cloud.
   http://www.isoc-ny.org/?p=1338