I'm in the tenuous position of trying to re-learn MPLS and BGP for an actual (largely cisco based) carrier network, as well as evaluate how FRR or other OSS projects might be integrated in the future. Doing these in parallel is... challenging :) Looking at the docs on https://frrouting.org/ it's not *exactly* clear what the current state is. I notice in the tests for LDP, there's a number of "all-fails", but I'm not familiar enough w/ the protocol to know whether those are strictly necessary for a working implementation. I was going to add that I wasn't sure if the "vpnv4" address family was supported but I found this <https://frrouting.org/user-guide/BGP-Address-Family.html#index-show-ip-bgp-vpnv4-all>. It explicitly calls out the supported RFCs, so I guess we're good there (though it should probably be noted in the "supported RFCs" page. If the contribution process is simple enough I'll probably submit a pull request). So the BGP support seems to be there, and LDP seems "mostly" complete, whatever that means. Does that mean FRR can actually integrate with a functioning MPLS network? How "production ready" should this be considered? Similarly, though I know this isn't exactly the responsibility of FRR, can anyone give me an idea of the state of open hardware support for MPLS?
From what I've seen so far, it's actually rather lacking, but I'm not confident I'm looking in the right places. Last thing I want to do is add an MPLS node that's going to punt every MPLS frame to CPU or something :)
Thanks in advance -- William R. George