Re: Private RANDFILE per CA required?

2012-10-04 Thread Stefan H. Holek
On 02.10.2012, at 15:22, Jakob Bohm wrote:

 On 10/2/2012 2:04 PM, Stefan H. Holek wrote:
 When using the openssl command line utility, is a private RANDFILE per CA 
 required for security reasons, or is it just fine to use a single RANDFILE 
 for everything (i.e. the default ~/.rnd)? Older configuration files seem to 
 indicate the former, but is this still true?
 
 IOW, I am looking for an answer to whether not having its own RANDFILE 
 degrades the security of a CA.
 
 I would say it degrades it, as it makes the randomness used by each CA less 
 random.
 
 I would also suggest getting a real hardware RNG source and directly or
 indirectly feeding it into OpenSSL.

Thank you for the answer. I am after something more specific though:

The openssl req, ca, etc. commands always load the RANDFILE, even if an 
acceptable source for seeding the PRNG exists. This means that RANDFILE is 
mixed into an already good seed [1]. Given that RANDFILE contains good data as 
well, I would assume this has exactly zero effect on the quality of the seed. 
I lack the maths degree to be certain about this though.

Thanks again,
Stefan

[1] http://www.openssl.org/docs/crypto/RAND_add.html

-- 
Stefan H. Holek
ste...@epy.co.at

__
OpenSSL Project http://www.openssl.org
User Support Mailing Listopenssl-users@openssl.org
Automated List Manager   majord...@openssl.org


Re: Private RANDFILE per CA required?

2012-10-02 Thread Jakob Bohm

On 10/2/2012 2:04 PM, Stefan H. Holek wrote:

Hi All,

Here is something I am not able to figure out, even after checking the FAQ and 
the rand(3) man page:

When using the openssl command line utility, is a private RANDFILE per CA 
required for security reasons, or is it just fine to use a single RANDFILE for 
everything (i.e. the default ~/.rnd)? Older configuration files seem to 
indicate the former, but is this still true?

IOW, I am looking for an answer to whether not having its own RANDFILE degrades 
the security of a CA.

Thank you,
Stefan



I would say it degrades it, as it makes the randomness used by each CA 
less random.


I would also suggest getting a real hardware RNG source and directly or
indirectly feeding it into OpenSSL.  These are commercially available 
and typically cheap, and they are still not included with most computer

hardware, despite many failed initiatives in the past.

Enjoy

Jakob
--
Jakob Bohm, CIO, Partner, WiseMo A/S.  http://www.wisemo.com
Transformervej 29, 2730 Herlev, Denmark.  Direct +45 31 13 16 10
This public discussion message is non-binding and may contain errors.
WiseMo - Remote Service Management for PCs, Phones and Embedded
__
OpenSSL Project http://www.openssl.org
User Support Mailing Listopenssl-users@openssl.org
Automated List Manager   majord...@openssl.org