Absolutely correct, but that just means that there has to be server-side locks for the commitment phase (4-6), it doesn't impact the client-side. This has an implication for performance of the commitment phase, but luckily, database vendors have been struggling with efficient implementation of commit for years, so using the transaction features of whatever database is used for storage should resolve most of those problems (i.e. check and update the version number in the database in a single transaction).

I agree the question doesn't exactly suggest a deep insight into networking, but since the submitter actually says so, I don't see the problem. As for wether such a question should make it to the front page, I actually think the question is interesting, at least from a computer science point of view.

As I see it, the only reason for having large backbones, in the terminology of the question, is to simplify routing. Thus, the only reason for having them, is the lack of a fast, world-wide, precise (as in up-to-date) routing algorithm, which could support a world wide network with millions of nodes. Recently, a lot of work have been done in routing protocols for large mesh networks, and the real question is: How large a network can be supported using similar techniques.

So i guess the answer to the question is: This could work with current hardware, if only we could implement a proper routing algorithm, which, most likely, we can't for a network the size of the Internet.