Re: [TLS] Impact of draft-ietf-mptcp-api on TLS

[[e-mail] (Martin Rex)]

Stephen Farrell wrote:
> 
>> 
>> I think the "since" part is wrong.  The security of the subsequent
>> subflows must depend on TLS itself, not MPTCP.
> 
> Fair point. I can try re-word it, but first to the real
> question I'd like to check...
> 
> Does anyone care that e.g. a server cert that contains IP
> address 'a' might be used for the first subflow, but
> subsequent packets for that TLS session can end up going
> vis MPTCP to IP address 'b' where 'b' is not mentioned
> in the cert?
> 
> Ah crap - I just remembered excludedSubtrees can apply to
> IP address name constraints. Anyone here care about that?


I somehow fail to understand the discussion.  The IP-address
that the socket shows after connection establishment is
completely irrelevant to the security of TLS, even when
Server-Certificates with IP-Addresses in the SubjectAltName
are used for server endpoint identification according to
rfc2818 section 3.1.

What is actually matched is the "information SOURCE" that the application
uses to establish the network connection and the Certificate presented
by the server in the TLS handshake.  Whether there are HTTP CONNECT
proxies or even network-address-transparent MITMs on the path does
not matter for the security provided by TLS, because that information
is not (must not be) part of the information that is matched
during server endpoint identification.

-Martin