> I would use multipath quic rather than tcp. And for bier CTL, we want a fine > hoppy protocol like PGMcc
Just one point about TCP vs multicast. You *can* get more paralellism from multicast where TCP from one central point (like a iBGP route reflector) will cause serialization delays. One could argue you get this with multicast by just moving the problem to the router that is replicating packets across an internal-crossbar switch or fabric. But you can imagine this is done by orders of magnitude less serializtion time. And your Sally reference to SRM is a really simple way to understand the concepts and how they *could* be applied to routing protocols. In the 90s, at cisco, we looked at replacing the iBGP mesh (read: mess) with SRM, or SRM-like mechanisms. It was pre-PGM days. Dino
