Re: [Mud] [Iot-onboarding] some straw-man charter text for an IoT Operational Security WG

[Michael Richardson <[e-mail]>]

Kent: I am feeling a bit of "my protocol vs your protocol" in your reply.
      It is not my intention to do that.  I think that we did an awesome
      job getting to the commonalities with 8366.  I'm not sure that this
      is as visible to others.

Kent Watsen <[e-mail]> wrote:
    > I'm still not 100% sure if this working group is needed but, if such a
    > working group is to be formed, I object to it being BRSKI-oriented, as
    > already there is ANIMA for that.

Yes, but is that really appropriate?  Are the right people there in ANIMA,
and is it actually disruptive to the ANIMA effort beyond onboarding?
In particular the constrained document is not getting enough attention.

    > I appreciate that the first paragraph
    > enumerates a number of efforts (BTW, it should be just "SZTP", without
    > the NETCONF qualifier),

Sure.  I put it there so that people could find it.

    > and later the text says that it's "not limited
    > to just that protocol" (being BRSKI), but the general tone, and
    > specifically the Milestones, are very BRSKI-specific.

I would very much like to have more, but I don't know what the next steps for
SZTP are.

    > To be clear, if
    > such a working group were formed, I would like to submit a document
    > entitled something like "SZTP for IoT and other Constrained
    > Devices".
    > In this document I could show how SZTP could be extended to
    > use DTLS, CBOR, and CoAP and further, that it achieves bootstrap state
    > with fewer message exchanges and greater flexibility, not to mention
    > that it already supports the offline/cloudless use-case and a
    > peusdo-NOOB approach (already implemented by Juniper).
    
I actually rather believe that SZTP is probably a better starting point for a
cloudless solution.
If SZTP can do a better job for constrained environments, it is *precisely*
this kind of stuff that I'd like to have in this group.

I'm okay with the WG producing 3 or 4 mechanisms, (and then, of course, an
n+1th in four years to converge them, applying
https://www.explainxkcd.com/wiki/index.php/927:_Standards)
Some have attributed this method to Scott Bradner, imaging a meadow of
many wild flowers.

-- 
]               Never tell me the odds!                 | ipv6 mesh networks [ 
]   Michael Richardson, Sandelman Software Works        | network architect  [ 
]     [e-mail]  http://www.sandelman.ca/        |   ruby on rails    [