One of the good things about the blobs protocol being so special is that it can be replaced with something else. The only thing the current messages do is reference a hash. Where you get that hash from doesn't really matter. So the more we can experiment with this the better. I'm also not super fond of the current blob protocol.
@farewellutopia-dev I wonder what the requirements are. Is this a replacement just for < 5 MB blobs? I guess we want to know better who was sending what, instead of doing just flooding. But maybe that is separate from moving the bits around? Hyper has been integrated with the SSB eco system before, so that is one candidate. We don't need to use the whole peer discovery part, we already have that.
Anyway, keen to hear what you think the minimal requirements are?