I have a hard time seeing how any sort of white list solution will actually mitigate any of the bad behavior exhibited by WoSign. From my perspective, I think we can make a pretty clear case that WoSign is a poorly run CA and poses a threat to the secure Internet that many of us are trying to achieve. They have many, serious bugs in their systems. Their responsiveness in fixing these problems is slow. Their understanding of security threats is limited. Their interest in compliance seems minimal. Their willingness to be forthright, honest, open in this forum can only be described as unacceptable.
So the problem I have with a white list is the implication that while we don't trust the CA to issue new certs, we do have trust in the continued operation of other parts of the CA. Chief among these is revocation as cert holders move away from WoSign to a new CA. Do we trust that WoSign will honor requsts for certs to be revoked? Do we trust that revocation will take place in a timely matter? Do we trust that WoSign will not collect information on hits to any OCSP responders they have set up and share that info with...whomever?
I'm just having a hard time seeing how there is anything left to trust when it comes to WoSign. Maybe the best outcome would be a finding of irreconcilable differences and for us to go our separate ways? Maybe we just want different things in a global PKI system?
On Wed, Sep 21, 2016 at 2:21 AM, Gervase Markham <g...@mozilla.org> wrote:
> On 12/09/16 22:46, Ryan Sleevi wrote:
>> Consider if we start with the list of certificates issued by StartCom
>> and WoSign, assuming the two are the same party (as all reasonable
>> evidence suggests). Extract the subjectAltName from every one of
>> these certificates, and then compare against the Alexa Top 1M. This
>> yields more than 60K certificates, at 1920K in a 'naive' whitelist.
>> However, if you compare based on base domain (as it appears in
>> Alexa), you end up with 18,763 unique names, with a much better
>> compressibility. For example, when compared with Chrome's Public
>> Suffix List DAFSA implementation (as one such compressed data
>> structure implementation), this ends up occupying 126K of storage.
> Can you tell us how many unique base domains (PSL+1) there are across
> WoSign and StartCom's entire certificate corpus, and what that might
> look like as a DAFSA?
I'm not sure about a DAFSA, but I wrote a semi-naive implementation of
a compressed trie and got 1592272 bytes. That is assuming each issuer
has its own trie. It could be optimized to be smaller if it was just
a single trie of eTLD+1 for all issuers.
dev-security-policy mailing list
_______________________________________________ dev-security-policy mailing list firstname.lastname@example.org https://lists.mozilla.org/listinfo/dev-security-policy