I created this SWaN using NetLogo complexity modelling tool. By rewiring just 8 links out of a hundred we got close to an optimal small world for this instance. Clustering is still dense relative to a much reduced average path length. This rewiring means normally distant connections have been brought closer to home. This is a model. In a Bitcoin network small world there would be a lot more clusters less densely packed I expect. But you get the idea. The question is, will the Bitcoin network form into a small world if left to its own devices - the intensely competing commercial operators of nodes.
This is a commercial problem, not a technical one. Not for miners, but for the Internet Exchanges Is the expected future Bitcoin business big enough and reliable enough for the Internet exchanges to adopt multicast? Multicast is not supported over the Internet. Period! Because it was never needed until now $ wise. It's technically simple. Turn on the switch and router software features which make it happen on intranets everywhere. Internet exchanges would do it if the business were there, like any business would. So this is a build it and they will come fallacy. There is no demand. Except for Bitcoin which also has yet to be adopted. So it's a quadratic fallacy. The challenge is to build a business case big enough and convincing enough for all the Internet exchanges in the world to start supporting it finally, en masse. It's a commercial problem, not a technical one. The alternative is to hash together a new host centric protocol that does it all, while scaling bigger than