Policy-Discussion

[Guillaume ROMAGNY <guillaume AT cacert.org>]

Hello Faramir,

Le 04/11/2011 01:51, Faramir a écrit :

El 03-11-2011 17:58, ianG escribió:
> following revocation of a sub CA in Malaysia is due to discovery
> of 512bit certs being used in an attack in the wild.  Some
> conjecture that the key may have been crunched.

  On 2009, RSA signing key for TI 83 calculator was factored on a
desktop computer, it took 73 days. Considering computers are more
powerful now, and malicious attackers might have larger budget to
mount distributed attack, I think RSA 512 re clearly weak. Maybe we
should not allow the usage of keys shorter than 1024 bits to sign
software.

About the TI factored key, you can read more at
http://en.wikipedia.org/wiki/Texas_Instruments_signing_key_controversy

+1

an email can be valuable for 1 minute to hours or days.
But software signing is meant to be strong enough for months.
So I guess you are right for codesigning.
At the same time, it could lead to more complex implementation. So unless the software team wants to make its life difficult, I would propose to switch to 1024 bit RSA key for all as a minimum

my 2 cents

--
Cordialement, Best regards,
Guillaume

Vision without action is a daydream.
Action without vision is a nightmare.  -- Japanese Proverb

Attachment: smime.p7s
Description: Signature cryptographique S/MIME