Re: [saag] A case against algorithm agility (long)

[Paul Lambert <[e-mail]>]

>
>> On 4/05/2014 04:10 am, Benjamin Kaduk wrote:
>> 
>>> That there is even a
>>> commercial market for it suggests that they are in use, and not just by
>>> people who could by pass the crypto.
>> 
>> 
>> Yes, this is fascinating.  There's also the SSL-interceptor boxes which
>> allegedly will take your commercially provided sub-Root.  Who uses those
>> machines and why?
>
>These boxes are used for whatever purpose people want access to the
>plaintext for. Uses range from running so-called "next generation
>firewall²([1]) policy on them, to looking for dissidents in certain
>countries.
>
>Commercial CAs rarely give sub-root certificates, and when they do, they
>have  name constraints ([2]), so they can¹t be used for generating fake
>certificates for mail.google.com.

No, the intermediate certificates sometimes do not have constraints. There
is a medium sized country that issued unconstrained intermediate
certificates to all of their middle school districts.  Also, revocation
checking for intermediates is typically broken in implementations -
particularly for OCSP stapling.

Paul


>These boxes generally either generate their own self-signed CA, or get a
>corporate sub-CA. Either way, it requires clients to be configured with
>the interception CA.
>
>Yoav 
>
>[1] ³next generation firewall² is a marketing term, but in general
>classic firewall can block addresses, protocols and ports, whereas ³next
>generation² looks at higher layers as well. So a classic firewall can
>block your access to Facebook, a next-generation firewall can block your
>access to Farmville.
>[2] Yes, there were cases where this was not followed.
>_______________________________________________
>saag mailing list
>[e-mail]
>https://www.ietf.org/mailman/listinfo/saag