Re: [Asrg] New proposal for spam blocking: Greylisting

At 1:24 PM -0500 6/20/03, Evan Harris wrote:
>  > > I would agree that users should be made aware of any possible 
>privacy risks,
>>  > but most users are already aware that mailer level statistics 
>>are kept, and
>>  > this is just the same thing, just a different format.
>>
>>  That mailer logs are also quite sensitive does not make your records
>>  less sensitive.
>
>The only issue I had with the original statement was the claim that this
>method should cause greater concern for privacy.  While there are privacy
>concerns with any statistics, I wouldn't consider this method any more of a
>danger than normal server logs, since nothing is kept that couldn't be found
>there.
>
>Evan

FWIW, my original point was that the grey list approach is a greater 
cause for concern than a simple white list approach (which can reveal 
a user's correspondents but not the date and time of the first and 
last e-mail and the number of successful e-mails).

I agree that the grey  list approach is not radically more 
problematic than mail server logs.  But even on that point, having 
two places where private info aggregates is worse than having one.

Moreover, the greylist might in some cases be easier to locate or 
access -- in some situations, for example a civil litigant might have 
fairly direct access to his or her grey list (and thus be easily 
required to produce it), but the related mail server logs might be 
held by a third party ISP not under the litigant's control.  And, 
once produced, the greylist can serve as an incomplete but still 
useful index into the mail server logs.

By pressing these points, I am not intending to be negative toward 
the proposal -- it certainly looks to be worth exploring further. 
But, when stacking this up against some of the other ideas and 
proposals, I think it is unavoidable that this has greater privacy 
implications than some other approaches.  Now if it is also 
significantly more effective, then greylists may well be worth the 
greater privacy risk (and there may be techniques that can reduce the 
privacy risks, such as the use of one way hashes of the triplets).

John

_______________________________________________
Asrg mailing list
Asrg@ietf.org
https://www1.ietf.org/mailman/listinfo/asrg