I suspected a distributed "server(? what would the proper word be?)" architecture from the image -- WONDERFUL concept WONDERFUL approach. THANK YOU in advance.
Its more likely serverless in a sense. I'm envisioning something X-like, where each one of us runs our own local "server" that handles and processes messages, and an XMPP-based communications architecture where players can connect to a central server that acts as a message dispatch which broadcasts all messages to all other local servers. In effect anyone running a copy of the message dispatch server is running a communications hub that allows game clients (through intermediate software) to connect to each other, and see what the others are doing.
Because the only centralized component is the message server, anyone running a copy can claim to not be running reverse engineered copies of the CoH server software, and can claim to not be infringing on any of NCSoft's content copyrights - because it doesn't contain any content. And because the system uses its own messaging system, it can't infringe on any of the technical APIs or interfaces of the game, in particular it cannot infringe on the (possibly protected) network code of the game servers.
The problem is that in this architecture, things like actual combat are problematic. The game clients have no awareness of how combat works, so they cannot be told things like "hit this entity for 200 points of damage" or anything like that. They won't on their own regenerate health bars or burn or recover endurance. They have no possible way of recording, much less understanding, what happens when one entity grants a power to another. All those things came exclusively from the game servers, and there's no "fragment" of them in the game clients Codewalker could leverage (to oversimplify a bit).
However, hypothetically speaking someone could actually write a combat engine and slip it into the system. If Codewalker was hit by a bus, a lot of his technical wizardry might get smooshed with him. But if his XMPP protocol is well documented (or even poorly documented but unencrypted and well-structured), anyone else could step in and write one that acts as a kind of "plugin" to the local message processors. In effect, each player would have to have a local "mapserver" that had the intelligence to perform combat, and it would need to be synchronized with all the others. You could also do this via an extension to the central server software which would eliminate the need for enhanced synchronization. The point being, you no longer need someone who understands how the game clients and the game servers worked on a very low level. You only need someone that could replicate the game functionality and understood how to add it to Codewalker's software instead.
Separate from the fact this almost certainly makes life easier for the SCORE team because they would be working in an abstracted environment much simpler than the raw one, it also opens the door for experimentation by the players.
Anyone with only knowledge of XMPP/XML and knowledge of how Codewalker's code works could theoretically experiment with extensions. Which is not to say I would rather have the effort be in any other hands besides Codewalker, but knowing him I suspect its intentional that one of the properties of the project is that the future does not lie exclusively in any one person's (or group's) hands. Once released, it can't be killed by any one person, or stopped by any one group. In fact, eight people who want to use the software can't be stopped by anyone, not Codewalker, not Titan, not NCSoft, short of physically going to their houses and taking their computers.
Just before shutdown, I wrote a long article talking about how a distributed MMO might work. One of the reasons why Codewalker's teaser seemed so obvious to me is that this is ground I've already traveled. Not that I'm taking any credit for any of this work: I'm certain Codewalker (and the SCORE team) is plenty smart enough to come up with this all on his own. Instead, I'm intrigued by the notion that its possible a conjecture I made a couple of years ago that I never thought would ever be tested might one day be tested: that its actually possible to make a game like this work.