> I'm sorry, to me this seems a bit farfetched to blame this case on HNCP.

Steven, nobody is speaking about blaming HNCP.

I am asking for a new feature that makes it possible for a network
administrator to easily disable HNCP on a given link.  I don't see how
that can be achieved otherwise.

If it can be achieved without changes to HNCP, please let me know how.

> to add logic and complexity to address this in HNCP

It's 20 lines of code, Steven.  The KILLER TLV has the following format:

   0                   1                   2                   3
   0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |            Type               |           Length              |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                         Milliseconds                          |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

If a KILLER TLV is received on a given interface, then this interface MUST
be put into external mode for at least the given number of Milliseconds.
If the interface is in authenticated mode, then the KILLER TLV MAY be
ignored unless it was received over unicast with proper authentication.

The idea being that by default all Homenet routers obey unauthenticated
KILLERs, but that they can be configured to ignore unauthenticated ones.
An hncp-killer works by sending periodic multicast KILLERs, and optionally
by sending unicast KILLERs upon receiving a NETWORK-STATE.

-- Juliusz

_______________________________________________
homenet mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/homenet

Reply via email to