[Madinas] my ideas about three documents

[Michael Richardson <[e-mail]>]

Our charter says:
} The MADINAS Working Group will document the current RCM state of affairs by :

} (i) identifying relevant network and application services scenarios and
} examining the effect of RCM schemes on them;

} (ii) analyzing various existing identifiers (i.e., beyond the MAC address)
} that can be used by the network to provide seamless services, and

} (iii) identifying scenarios where device identity is not required.

} The group will generate a Best Current Practices (BCP) document
} recommending means to reduce the impact of RCM on the documented use cases
} while ensuring that the privacy achieved with RCM is not compromised. For
} scenarios where device identity stability is desirable, the BCP document will
} recommend existing protocols that can be used to protect the request and
} exchange of identifiers between the client and the service provider.


To my mind, the first document should be a document that explains what RCM
is, and how we can talk about it.    To that end,
  https://www.ietf.org/archive/id/draft-ietf-madinas-mac-address-randomization-06.html
is close to the right document, and it's why I wanted to put the taxonomy in.
This document is about the *devices* and what they do, not about the
networks.   There is some additional content that needs to go in about
Pre-Association addresses vs Post-Association addresses.  There are some
hardware limitations that may exist in ability to change addresses at various
points.  For instance, can one do EAP/WPA activities on one address, and then
once the end host is convinced it's a network it wishes to join, switch to
another address.


The second document, draft-ietf-madinas-use-cases-05
Randomized and Changing MAC Address Use Cases and Requirements
is a bit further from what I have in mind.  Im particular, I would like to drop
"Requirements" from the title.
It's also unclear if these are reasons for RCM ("RCM Use cases"), or cases
where the network infrastructure has used MAC addresses as identifiers.
The way that I'd like to re-structure this document is to import much of the
excellent work that Amelia has done at:
   > https://mentor.ieee.org/802.11/dcn/21/11-21-0332-37-00bh-issues-tracking.docx

and then amend each case with an analysis of what happens when the different
policies (the PBGM, PNGM, etc.) occur in those cases.  For instance, all
policies except for PPGM work fine for the airport security line effort.
But, devices that always pick a new address for pre-association
communications, would completely break such a thing.
This document is not about RCM, but rather is about networks which have
assumed non-RCM.

The third document, the BCP, needs to heavily reference the second document.
In this document, we need to see if changes to technology used (protocol
options) the various use cases from document two.
There are various proposals that I'm unfamiliar with from the IEEE, and maybe
even the WBA, about things that could be changed at layer-two.  And probably
there are some improvements that could be made.  For instance, a four-address
mode that allowed the inner addresses to be encrypted easily.

My personal view is that this is an opportunity to upgrade the quality of the
identities used in devices, to finally create a ecosystem where the UX around
client certificates does not suck, and to get rid of the very insecure
network PSK.

--
Michael Richardson <[e-mail]>   . o O ( IPv6 IøT consulting )
           Sandelman Software Works Inc, Ottawa and Worldwide