[anonsec] BTNS C-bindings slides

[Miika Komu]

Hi all,

below you can find an updated version of the BTNS C-bindings slides for
IETF68. Unfortunately, we had an older version of the slides in the WG
meeting. I had added a proper intro to the new slides, but otherwise they
are the same:

http://www.iki.fi/miika/docs/btns-api-ietf68.pdf

[Nicolas Williams]

Post by Miika Komu
http://www.iki.fi/miika/docs/btns-api-ietf68.pdf

BTW, channel bindings are best thought of as a property of pTokens, not
iTokens.

Unique channel bindings (e.g., if we go the SKEYSEED route) can only be
associated with specific channel instances, whereas end-point channel
bindings can be associated with either the channel or the end-point
identities. Also, we need to apply a small transformation to the
end-point public keys in order to produce a channel binding octet string
(see other posts from Julien, Michael). Therefore it's best to always
think of channel bindings as an attribute of a channel.

Also, a peer identity (iToken) might have multiple names (though only
one asserted name); its asserted name will never be a public key unless
the peer is a BTNS peer.

We do, however, want the peer public keys, if there are any, to be
available through the iToken.

Recap:

channel binding octet string -> attribute of pToken

end-point public keys values -> attribute of iToken

Nico
--