Re: [secdir] Secdir review of draft-ietf-mpls-tp-mip-mep-map-09

["Adrian Farrel" <[e-mail]>]

Hi Vincent,

Thanks for your time.

> This document refers to [RFC6371] and [RFC6941] for detailed security
> discussions. I have no problem with that. However I have two comments:
> 
> 1/ It says:
>    "Implementations therefore are required to offer security mechanisms
>     for OAM.  Deployments are strongly advised to use such mechanisms."
> These sentences do not use the RFC2119 key words. Is that deliberate?

Yes.
This is an Informational and Design Considerations document, not a protocol
spec.
The normal English language seemed perfectly fine for conveying the consensus of
the WG.

> 2/ I really have problems understanding the following claim:
>    "Mixing of per-node and per-interface OAM on a single node is not
>    advised as OAM message leakage could be the result."
> Can you be more explicit in the I-D? It's important since it's probably not
> discussed in [RFC6371] and [RFC6941].

It seemed to the authors (and the WG - in fact, I recall this text coming from
the WG review) that an implementation that attempts to handle per interface OAM
and per node OAM is almost certain to get confused and send the wrong OAM
messages at the wrong time.

I would not go as far as to suggest that the quoted text is particularly
important, and the only way I can think to clarify it would be by adding "unless
an implementation is particularly well designed and tested."

> Minor comments:
> 
> ** MEP is used without being defined.

You're right! 
Amazed that it slipped through. Thanks.

Cheers,
Adrian