On Tuesday, April 28, 2020 5:07:00 PM CEST Fernando Gont wrote:
While looking at batman, I came across your IETF Internet-Draft
1) Is this the closest there is to an specification of batman?
Yes, that's the closest complete specification. There are various aspects
documented on open-mesh.org
in the documentation section, although those are
not really complete.
2) Does it described the current protocol, or have there been changes
since then that have not been reflected into the internet-draft?
There have been various changes. The algorithm described in the internet draft
was called B.A.T.M.A.N. III ; there have been amendments which have been then
implemented as algorithm version B.A.T.M.A.N. IV. Since then, batman-adv has
been implemented as Layer 2 mesh software (BATMAN as described was a Layer 3
routing daemon). There have been more modifications since then like multi
interface routing, bonding, interface alternating, bridge loop avoidance etc.
3) Any reason why the draft wasn't published as an IETF RFC?
The initial draft received various devastating reviews from some professors,
including missing mathematical proofs and that we should change the name since
it was considered ridiculous. We thought otherwise, but didn't want to spend
energy making reviewers happy. Instead we focused on the software itself. I
believe the original draft received around a thousand scientific references
(although nobody should reference IETF drafts, but after all its the most
complete document there is). Batman-adv is now in the Linux kernel, and the
old Layer 3 BATMAN daemon development is pretty much abandoned for years, but
still available for download. Almost everyone is using the layer 2 variant,
which wouldn't fit well with IETF in the first place but would more be an IEEE