On Tue, May 29, 2018 at 06:00:12PM +0200, Daniel Cegiełka wrote:
> Hi,
> 
> Sorry if this is a more "misc-list" topic, but according to
> Cryptography Services[1] team:
> 
> "SPHINCS[2] is the more recent one (vs XMSS), combining a good numbers
> of advances in the field and even more! Bringing the statelessness we
> were all waiting for."
> 
> Would SPHINCS be not a better choice then XMSS?
> 
> [1] https://cryptoservices.github.io/quantum/2015/12/08/XMSS-and-SPHINCS.html
> [2] https://sphincs.cr.yp.to/index.html
> 
> Best regards,
> Daniel
> 

Hi,

first of all:  This is an old comparison.  SPHINCS and XMSS have both
evolved since 2015.  SPHINCS [0] got superseeded by SPHINCS+ [1].

SPHINCS had 41kB sized signatures, SPHINCS+ can go down to 8 kB if you
are using lowest security options, like NIST defined in their process.
XMSS usually creates even smaller signatures, if that is an important
factor for you.

The reason SPHINCS+ can be stateless is that it uses probability.  The
signatures are created using Few-Time-Signature schemes.  You can sign
with the same key pair a few times, but the security decreases with
every signature.  Using the combination of a huge tree and that scheme
you can show that (with a given set of parameters) the probability of
re-using the same key pair too often decreases.  On XMSS you keep state
so that you definitely not re-use the same key pair.

The most important factor for XMSS is that it is an RFC, RFC 8391 [2] to
be precise.  It has been through a standardization process including
proper review and comments.  SPHINCS+ on the other hand is in the NIST
Post-Quantum standardization process, but it will probably take a few
more years until it completes.

Patrick

[0] http://sphincs.cr.yp.to
[1] http://sphincs.org
[2] https://tools.ietf.org/html/rfc8391

Reply via email to