NOTE: you are reading this on NDNL. The new blog is now at http://nilsgeylen.com/blog and you should visit it.
Cloud computing is an often used term these days. But what is the cloud precisely? Who owns it? How do we access it? And more importantly, are we using it as we should? As far as I’m concerned, I think we’ve been doing it all wrong. And now it seems Windows will come to the rescue.
According to Wikipedia, cloud computing is to be understood as
a style of computing where IT-related capabilities are provided ‘as a service’ allowing users to access technology-enabled services from the Internet (‘in the cloud’)
While that definition remains vague, many of us are using cloud services already.
Typically, we think of the cloud as the combination of (free) services that you use to store and manage your data, such as Gmail or Flickr. After all, the cloud is considered a synonym for the internet.
Technically, however, the cloud could be anything from Google data centres to your own host. But even the ISP which stores your emails and your own machine can be considered part of the cloud: whenever you sync your email to Gmail, when you use P2P networks, when you post crash reports or usage stats (e.g. scrobbling to Last.fm or Wakoopa) you are, in a way, contributing to the cloud.
4 ways to the clouds
SaaS
The term Software as a Service (SaaS) is often used as a synonym for cloud computing (although strictly speaking it isn’t). It implies the ad hoc access to online tools that work as applications. Of course, you will always need another application to access the tool. Google Docs for word processing is an example of that: you do not need to install Word to write a letter – but you do need a browser to access your Google account.
RIA
Rich internet applications are another way of accessing the cloud. In that case, you do not even need a browser. With Adobe’s AIR for instance, all you have to do is install the framework, get the app you need, and you’re set. An AIR app runs on your desktop and communicates with your online service of choice. Twhirl for Twitter is an example of that.
Widgets
Much like RIAs, widgets run on your desktop. Unlike RIAs they do not strictly need a framework. With Vista or OS X they run natively within your operating system. Yahoo! Widgets works in a similar way, although you still need to install the Yahoo! widget engine (formerly Konfabulator) to run them. And Google has its own sidebar with widgets too. In a way, widgets are just skinned miniature apps.
Software with services
A final way of accessing the cloud is one that sounds very new and very vague, but promising as well. In fact it’s what Google is doing with Chrome and what Microsoft would like to see with Silverlight. The idea is not to use a separate framework (like Adobe’s AIR), but to present the content in a browser as a separate process.
This becomes most clear in Chrome, where you can unlock a service (say Gmail) and have it run in a separate application window, independent from the main browser process. Silverlight, as I understand it, is also a way to present rich internet content in the browser. So far, I haven’t seen any breakaway usages like with Chrome, but that may soon change.
Current rumours about Windows 7 will have us believe that the successor to Vista will not feature any pre-installed, standalone applications for email or multimedia (read: Mail and Media Player). Instead, Microsoft is looking at a more enhanced way of using their online counterparts: Windows Live and Office Live and Silverlight.
And now, apparently, Microsoft is about to introduce yet another, newer incarnation of Windows – aptly codenamed Cloud. This mystery product seems to involve “software plus services where web-based components supplement functionality of the main desktop software”.
Personally, I think that is in fact the way to go.
The fog
It’s clear that the cloud is everywhere and nowhere. Like the internet, no one owns it. That implies the data is yours… but the service provider owns the tools. In the end, that approach locks you in as much as any proprietary desktop software package would have: if the service changes or collapses, your data is gone.
Furthermore, at any given moment I may have some documents with Google, my mindmaps could be with MindMeister and my wiki with MindTouch. That way, the cloud becomes a thick, dense fog. And if I’m having trouble managing my own stuff, what about people trying to connect with me, using other services and frameworks?
Think: if you want to share your data in the cloud, you have to invite collaborators, viewers, etc. Now you’re not only locked into the service, but into a maze of passwords and guest accounts and you’re not sure what happens next.
The solution
When setting up this new, self-hosted blog, I chose to have my lifestream self-hosted too. That way, I take back part of the cloud and I can share it as I see fit. I may set up a wiki at some point, and if there was some document sharing software, I’d use that too.
The way I see it, I need storage for my data, an on-the-fly ‘sotware interface’ to access it and an app to tie it together. That app could be a browser like chrome or perhaps this new Microsoft Cloud ‘software with services’.
That way, I would have my files on my server, use an online service to edit it, and have my desktop application as a gateway. But the files I edit are mine. They’re not locked in and they don’t die when Google does.
Of course, owning your part of the cloud, also means paying for it. But in my view, I think I’d rather pay for that ‘connective software interface’ than for yet another web 2.0 service that effectively owns my data (Flickr locks you out of downloading your own pics, for instance, when your account expires).
I may be all wrong about this, but one thing’s for sure: the cloud, we’re definitely not in it yet.