The Opera Unite vision video explains their take on the small cloud. Rather than our current two-tier system, with clients and servers occupying distinct roles, we can move to system that lets every computer fully participate in the web.

Clearly, Opera understands Small Cloud Theory.

Unite uses an architecture similar to the one we described in our series on Ending the Paper Shuffle last year. (See locating documents, versioning documents, and tracking documents for our thoughts on how to address common information handling problems in the enterprise.)

Coming soon: the future

Opera claims this approach “reinvents the web” but I think it simply delivers on the Internet’s original promise. The original HTTP drafts looked ahead to a more peer-centric approach to networking. You’ll rarely find the words “client” or “server” in W3C’s final recommendations; they prefer the terms “browser” and “host” — and implies that these roles are not exclusive.

Peer-to-peer technology is hardly new,although it’s been slow to shed its association with questionable file sharing practices. On the subject of P2P, Clay Shirky wrote in 2000:

The fusion of desktop and server… is turning the internet inside out.
The current network is built on a “content in the center” architecture.
However, the ability for a desktop machine to take on the work of a server increases annually.
…Add to these forces an increasing number of PCs in networked offices… and you have the outlines of a new “content at the edges” architecture … This is the future, and Microsoft knows it.

I’m not sure if Microsoft really “knew it” — although their efforts with acquiring Groove, and more recently LiveMesh might indicate that it’s possible that they did — but I do know this: The computer I’m writing this post on is about 22 times more powerful that the average web server in 1996. In addition, it has 285 times more available memory. And it’s not even a particularly fancy or fast computer — a mid-range Dell business laptop.

So although Shirky’s “Fusion” of the desktop and server isn’t quite here, 8 years later, Opera’s Unite might help bring it a step closer.

Related posts:

1. Small Cloud Theory
2. Enterprise 2.0: What’s up with Google?

If you’re in the Northern Virginia area and you’d like to stop by for a drink, we’d love to see you – onwards from 5:30PM.

View Larger Map

Related posts:

1. A View from the Infovark Burrow
2. Viral Marketing

We just missed an interesting sounding opening presentation from Rob Carter from FedEx – it looks like FedEx are making extensive use of the web, facebook and blogs and wikis both within and external to their organization.

Sean Dennehy and Don Burke then presented a great seession on their work at the Central Intelligence Agency. The CIA’s knowledge sharing ability has been greatly enhanced since they deployed their “intellipedia” – a mediawiki implementation that allows CIA staff to edit and share information freely, and without editorial regulation.

My favourite quote from Sean – “Wikis don’t work in theory – they only work in practice”

Other than removing the ability to make anonymous edits, not much was changed by the CIA when they launched intellipedia, last year.  They claim also to have a much higher contribution rate (Wikipedia has a markedly low percentage of users who actually edit it – often guessed at about 1-3%) – but they are still working with the early adopters – intellipedia hasn’t been wholly rolled out to the entire organization.

“A culture problem – not a technology problem”

Don mentioned that there was substantial resistance to their efforts to incorporate this crazy wiki thing into their business. Primary benefit comes from working at the broadest audience possible. The wiki approach also focuses more on topic than on organizational structure – it means that the point tends to be on content, rather than process. That’s a really good thing.

” But – I don’t have time to edit this intellipedia thing”

Don and Sean seem adamant that the best way to deal with this kind of response is for people to stop writing emails and documents, and start writing intellipedia articles instead. (I suspect that that’s going to be a friction point for them. People don’t like new ideas very much. )

All in all, this was a great session – intellipedia seems set to be a great success.

Related posts:

1. Find us at the Enterprise 2.0 Conference
2. Thoughts on the Viability of Enterprise 2.0 Webinar
3. Administrivia

When we analyze the cloud, we can see that while the term “The Internet” is often used as a singular noun, it is in fact,very plural. The Internet is made up of thousands of other networks. That’s what the word Internet means — a network of networks. The cloud is itself a cloud of clouds.

If we look at a map of the Internet, we see a cloud made up of thousands of connected clouds. Each cloud represents a cluster of IP addresses owned by various different organizations. And if we zoom in on those smaller clouds, we find a repeating pattern of nodes connected to other nodes. Some nodes are highly connected within the cloud; some nodes have just a few connections. Zooming in further, we eventually find web servers publishing information within particular organizations. And if we zoom in at the highest resolution, we find each of those web servers connected to still another cloud of nodes: the workstations that access those servers.

At each level of magnification, the Internet appears the same: It’s a loose collection of nodes, a few of which are highly connected hubs, but the vast majority of which are spokes with very few connections. In mathematical terms, it’s called a scale-free network.

Scale-free networks are a hot topic right now, and not only because of the Internet. Social networks are also believed to be scale-scale free networks, as is the human brain. Scale-free networks also exhibit some interesting properties, in that they are highly resistant to random failures but somewhat susceptible to epidemics. The distribution of highly connected nodes to nodes with few connections is a power law distribution, which generates that famous long tail graph.

Those small clouds we see at the lowest level of magnification are the main source of activity on the Internet. This is where all the work gets done. I’m typing this on a PC connected to the VarkNet in the infovark Burrow. You’re probably reading it on a PC that’s connected to your own organization’s network. All of the information on all of the web servers across the entire Internet is ultimately authored by people on personal computers connected via their local networks. Everything on the Big Cloud started on a Small Cloud.

Here’s another interesting thing about a scale-free network: Its rules operate the same way at a macro level as at the micro level. It’s fractal in nature. The same patterns hold, no matter at what resolution you view the system. That’s why the same technology that allowed a handful of academics to send doctoral theses to each other can scale to the point where millions of users can share video with each other. There’s no magical point where Small Cloud architecture becomes Big Cloud architecture. The Small Cloud is the Big Cloud.

The output of all our collective activity on the Internet is information. It’s policies, email discussions and contracts. It’s blog posts, and wiki pages, and accounting spreadsheets and customer service records. Where should that information belong?

As far as the Internet — the network of networks — is concerned, it’s all just a bunch of ones and zeros. It doesn’t really matter where the bits go. Anywhere in the Cloud will do. One network is just as good as another.

Related posts:

1. Opera Understands the Small Cloud
2. Legacy Theory
3. Information Hubs and Twitter