Bryght started with a simple idea: what if we could take a powerful, complex framework like Drupal and make it available to everyone... with or without technical expertise. This was closely related to our belief that eventually static HTML pages on the Internet will be replaced by dynamic pages. Dynamic pages means web applications.
We've seen a lot of this occur over the past several years. When we started, Drupal was making the transition from Drupal 4.4 to 4.5. Multisite was just a glimmer in people's eyes, and the concept of install profiles was nowhere to be seen. Bryght worked on Drupal core to include multisite capability out of the box: all of a sudden, it was a bug if a module didn't support operating in a multisite environment.
Drupal-as-framework was in a much different state back then. Developers constantly hit cases where the core code didn't have enough "hooks and interfaces" to cleanly override everything. We ended up building a series of tools and daemons around Drupal in order to enable mass hosting. This tool is called HostMaster, and is Bryght's answer to mass hosting Drupal. It's built around Python and PostgreSQL, and has had the concept of "install profiles" for about 2 years. We originally had dreams of perhaps licensing or otherwise making money directly off this code. But in reality, this concept is foreign to our open source beliefs: the bits don't matter. Eventually, we put HostMaster under the GPL and made it available at https://svn.bryght.com/hostmaster (yes, you still need to request an account).
Update: I had to repost this and rewrite this bit at the end (don't ask...). It does have the nice side effect of me being able to point back at Marc Canter: say what you want about Marc, he's been thinking in "network of networks" mode for years, and has been trying to get the rest of us to catch up with his thinking.
Fast forward to the eve of 2007, and the concept of multisite and mass hosting is everywhere. The folks at Wikia are thinking about Drupal mass hosting, and not a week goes by when we don't hear about other projects thinking about varieties of multiple sites. We're entering the stage of networks of networks.
Drupal is much more of a framework today, and is now ready to handle driving this kind of network itself. We're looking at rewriting some of our mass hosting tools on top of Drupal directly. Yes, I'm looking at you and your Provisionator stuff, D'Arcy Norman -- except instead of working within just a single shared hosting environment, we can look at maintaining sites in a distributed fashion: think XML-RPC and operations across sites, from a dashboard type of interface.
Of course, the next step is shared user profiles, personas, and identity. Drupal has had distributed authentication, and we've built profile sharing on top of that over two years ago, in the form of the FOAF and SXIP modules, but the emerging standard for user centric identity is OpenID, especially with v2 that includes attribute exchange courtesy of the original SXIP concept. But James Walker, author of the previous two modules, hasn't been sitting around just waiting: building on Rowan Kerr's work, we'll soon have a full OpenID server and client solution, making it easy for Drupal to be the hub or a leaf in a distributed social network.
Hooks and interfaces are everything: the standard for the identity piece, while not perfect, is solidifying. RSS is a standard transport for pushing around semi-structured data. XML-RPC can be used to programmatically get more communication between sites. XMPP is emerging as a standard for real-time message passing. All the pieces are coming together.