The Project M devs would have to do a complete overhaul of the actual online system that was grandfathered into Project M because it is a internet service that is UNABLE TO satisfy the needs of a fighting game, those needs being precision and fluidity (these are the only ones I can think of off the top of my head, feel free to throw in more). WiFi butchers these needs, guts them and eats them for dinner due to it's inherent input lag and internet speed being determined by the worst client.
The devs would have to figure out a way for the Wii to NOT connect to Nintendo's servers during play sessions of Project M, which is just a gargantuan task in itself since that crops into the boundaries of hacking your Wii to recognize and be able to connect to a second service provider. Next would be the creation of the new online system that Project M would use and for Brawl to be able to actually recognize this new game-play element, apply the new rules to its code, and work flawlessly. The networking architecture that I believe would work best for Project M is the one used by everything that ISN'T made Nintendo; the Client-Server architecture. As you may or may not know, setting up a functioning network
is not a simple feat (I know that this is referencing Source engines, but the idea is basically the same). Servers would have to be created and maintained; not just one server, but many, MANY servers to host the player-base of Project M, which at that point the devs might as well just make another game and charge for that since servers cost money to run. Not to mention handling the issues WiFi has infected the game with.
This is all hypothesis of course, but it should shed some light on the amount of time and resources that potentially could go into creating a working and playable network service.
