On 21/03/2013 07:47, Carl S. Gutekunst wrote:
On every SMTP server in the world, the first 100 to 250 bytes send by
the server after TLS negotiation completes are both constant and
known: they're the EHLO response. So that got me wondering if
randomizing the EHLO response could be helpful in mitigating
statistical attacks. Obviously the degree of "randomizing" is limited;
RFC 5321 requires the first line to be the FQDN of the receiver, and
every line will contain "\r\n\220-". The most I can do is randomize
whole lines. But I don't have an intuitive feel for what might be
effective in disrupting these statistical code-breaking techniques;
hence the question.
Isn't it '250-'?
Anyway, I'm not a cryptographer either, but my limited understanding
suggests that things like TLS work on a 'block' of data at once, not
byte-by-byte. So, if this is deemed a risk or potential risk, it would
make sense to make the first 'block' as random as possible. AIUI TLS
uses the previous block as part if the encryption for the next block, so
if they can't tell what the first block contains, they can't do the rest
of the stream either.
So, randomising line order, possibly adding 'x-messwithhackers: <random
string>' pseudo-capabilities, adding random text after the FQDN of the
receiver on the first line of the EHLO response, etc, would all stand a
fair chance of making it harder to hack.
But, I'm not a cryptographer, so I may be talking total garbage :-)
Paul Smith Computer Services
Tel: 01484 855800
Vat No: GB 685 6987 53
ietf-smtp mailing list