Re: [ietf-privacy] [Tzdist] [saag] Fwd: WGLC for draft-ietf-tzdist-service-05

[Paul Eggert <[e-mail]>]

Cyrus Daboo wrote:
> I will try and propose some text for a Privacy Considerations section

Although we will need such a section, the tzdist privacy issues dkg raised 
shouldn't be addressed merely by adding a section that says in essence, "yes, 
admittedly tzdist's privacy stinks, but at least we've clearly documented that 
it stinks".  Let's instead use his helpful review to adjust the protocol so that 
it better preserves privacy.

Doing that will take some work, but it's doable.  Here are a couple of thoughts 
in that direction.

First, we can prevent servers from tracking users via ETag or steganographic 
data by requiring standard ETags (e.g., "tz2015a") and normalized data.  Yes, 
this is a bit harder to implement on the server side, but it's not *that* hard, 
and it does prevent this privacy abuse.

Second, we can prevent tracking users via query parameters by making the typical 
query be simply "Give me everything", thus avoiding the parameters.  As the 
entire tz database can be communicated in 25 kB in compressed format, this will 
be reasonably efficient (when combined with standard ETags) and will better 
preserve privacy.  Yes, this also requires more work (on both client and server 
side), but it's not that much work, and if we really want privacy it should be 
work we're willing to do.