In article <[EMAIL PROTECTED]>,
Werner Koch <[EMAIL PROTECTED]> writes:
> Glynn Clements <[EMAIL PROTECTED]> writes:
>> A quick check seems to indicate that /dev/random won't be able to
>> provide enough data for generating ephemeral port numbers unless the
>> system is *very* lightly loaded.
> And please do not use /dev/urandom either - it would provide enough data
> but empties the random pool, so that /dev/random will practically always
> block while waiting for new entropy. You don't need cryptographic
> strong random numbers.
You would of course only use it to seed a pseudo random generator, perhaps
every 60s.
Unfortunately it would be still easy for an attacker to guess the state
based on a few samples, to be reasonably secure it would need to use a
similar algorithm as the initial TCP sequence number, which hashes the
current state with a timestamp and unique information (in this case the pid)
-- to prevent the attacker from guessing the state.
Unfortunately using such a hash is insanely slow. So overall I think
it is a bad idea. Naive "solutions" don�t cut and real solutions are too
slow. So I would just ignore it.
-Andi
> Werner Koch at guug.de www.gnupg.org keyid 621CC013
> -
> To unsubscribe from this list: send the line "unsubscribe linux-net" in
> the body of a message to [EMAIL PROTECTED]
-
To unsubscribe from this list: send the line "unsubscribe linux-net" in
the body of a message to [EMAIL PROTECTED]
