On 23/09/2016 14:12, Kurt Roeckx wrote:
On 2016-09-23 13:38, Richard Wang wrote:
Hi Gerv,
Please check this news (Feb 25th 2015) in OSCCA website:
http://www.oscca.gov.cn/News/201312/News_1254.htm that all China
licensed CA finished the PKI/CA system upgrade that all licensed CA
MUST be able to issue SM2 certificate to subscribers.
As I said in last year CABF face to face meeting in Switzerland,
WebTrust is USA standard, ESTI is Europe standard, I think China have
its own standard also. This a problem for global CA that have business
in worldwide countries that maybe need to setup many roots to manage
for complying with different standard.
We know issuing SM2 cert is not complied with BR, but you can treat it
as "compelled" by regulations, so we need to test the gateway
installed RSA certificate and SM2 certificate in the public Internet,
to test the auto-negotiation from browser to gateway, if the browser
like Firefox don't support SM2, then the gateway will use RSA
certificate for communication, if the browser like 360 browser that
support SM2, then use SM2 certificate.
There seem to be several governments that define their own standard,
like GOST in Russia, SEED in South Korea, and the SM2/SM3/SM4 in China.
I guess you could also see AES as a USA standard and Camellia as a
Japanese standard.
I think in this case SM2 should be compared to the NIST curves, whose
design was not reviewed by anyone from outside NIST/NSA. Outside the
CA/B forum, there seems to be growing support for other curves such as
the BrainPool curves and the Ed25519 curve.
Internationally we do not want to support all such standards, which is
why we select some. I think this selection is mostly based on the trust
that there is in that algorithm based on international review of them.
The only suggestion I have is that if the government requires you to use
those algorithm for certain certificates that you use a separate CA root
for that.
Such artificial fragmentation reduces algorithm and curve agility in
the worldwide Internet, increasing the risk that the only "permitted"
algorithms are all broken before replacements become "permitted".
having a specific BR rule banning any curve except 3 curves from a
single government project in a single country certainly looks like a
very bad idea.
Enjoy
Jakob
--
Jakob Bohm, CIO, Partner, WiseMo A/S. https://www.wisemo.com
Transformervej 29, 2860 Søborg, 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
_______________________________________________
dev-security-policy mailing list
dev-security-policy@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-security-policy
