Re: [keyassure] Objective: Restrictive versus Supplementary Models

[James Cloos <[e-mail]>]

>>>>> "MR" == Martin Rex <[e-mail]> writes:

MR> In a first round of attack, the attacker inserts a fake unsigned
MR> TLSA record (DNS poisoning) that the victim is accessing with TLS
MR> frequently and where the DNS admin is not using DNSSEC.

Why would an attacker insert a fake tlsa w/o a matching fake a or aaaa?

More likely they inject both and send the user off to a fake site with
matching fake credentials.

The client software might notice that the creds have changed since the
last time it visited (/if/ it has ever visited), but how many do that
for longer than a "session"?  Any?

Without dane the attacker just does the fake a/aaaa, fake server and
fake cert an still wins.  So it isn't a regression.  But dtls w/o
dnssec doesn't actually help either.

Injecting a dtls w/o also injecting other fake RRs might get done as
a lark, but any real attacks will be more sophisticated.

-JimC
-- 
James Cloos <[e-mail]>         OpenPGP: 1024D/ED7DAEA6