Re: [6tisch] Benjamin Kaduk's Discuss on draft-ietf-6tisch-msf-12: (with DISCUSS and COMMENT)

[Benjamin Kaduk <[e-mail]>]

Hi Mališa,

On Wed, Mar 25, 2020 at 10:37:57PM +0100, Mališa Vučinić wrote:
> Hi Ben,
> 
> There has been an extensive discussion on this issue in the WG. As Tengfei stated, since MSF operates exclusively at L2, reading DSCP values from the IPv6 header would constitute a layer violation. It was decided that MSF would implement the recommendation from draft-ietf-6tisch-minimal-security by recommending the rate limit on DSCP-tagged traffic, at IPv6 layer as outlined in Security Considerations. That said, other scheduling functions that may operate higher up in the stack, e.g. to establish end-to-end tracks between nodes in a mesh, may adhere to this requirement from minimal-security. Therefore, for the sake of future scheduling functions that may get standardized, it was deemed appropriate to leave the recommendation in minimal-security as-is.
> 
> Hope that clarifies.

It does help clarify, thanks.
I'm still not happy about ignoring the SHOULD from -minimal-security but
can't really refute the technical points being made, so I will remove the
Discuss point about it.

That said, it seems like a slight rewording of -minimal-security might be
in order, since:

   AF43.  Companion SF documents SHOULD specify how this recommended
   behavior is achieved.  One example is the 6TiSCH Minimal Scheduling
   Function [I-D.ietf-6tisch-msf].

seems to imply that MSF is an example of "specify[ing] how this recommended
behavior is achieved", since it is just left to "the implementation" (i.e.,
not fully specified).

Thanks,

Ben