Re: [babel] Mirja Kühlewind's Discuss on draft-ietf-babel-hmac-08: (with DISCUSS and COMMENT)

[Juliusz Chroboczek <[e-mail]>]

> You can also use a different timer e.g. 300s which would also ensure
> that a challenge is not send more often than 300s (expect in the case
> where you got a challenge reply but in case you are not talking to an
> attacker and usually not need to send another challenge request immediately).

If the challenge reply is lost, then the challenge needs to be resent.  It
would not be acceptable to cause a 5 min blackhole after a single packet
loss.

> However, in any case I find 300ms rather low. RFC8085 recommends
> basically one active message (in case you have a response-reply pattern)
> or not more than one message every 3 seconds.

This would mean creating a 3 second blackhole after a single packet loss.

I'll expand on that further in my reply to your review of 6126bis, but
I believe that RFC 8085 speaks about UDP traffic across the Internet.
Babel is a link-local protocol -- there are no intermediary nodes to
congest.

Mirja, if the RFC 8085 limit is to be enforced for link-local protocols,
then OSPF cannot possibly work.

-- Juliusz