Re: Email Web of Trust

Alan DeKok wrote:
> Yakov Shafranovich <[email protected]> wrote:
> 
>>What I am looking for is a one or two sentence problem statement for 
>>this concept, similar to what we have in the LMAP document:
>>
>>"  LMAP attacks the forgery problem by checking that the host
>>    from which the message was sent is authorized to send mail
>>    using the a domain in the message's envelope.
>>"
> 
> 
>   The web of trust attacks the spam problem by collating information
> across multiple recipients, to determine the probability that a
> particular SMTP sending IP is sending spam.  This probability may be
> calculated (with some level of confidence), even if a particular
> recipient has not had prior interaction with a particular sender.
> 

I would rather not restrict it to IPs, but rephrase "SMTP sending IP" as 
"MTAs". If somekind of LMAP-like scheme is implemented, it is possible 
that this web of trust can be used for domains as well.

I also don't like the phrase "attacks the spam problem". I would rather 
see something like "helps to fight the spam problem" since it makes it 
sound like a silver bullet although I could be wrong.

> 
>   It won't stop spam, but it will give people more information than
> they have today.
> 

Agree. We will have to determine at some point whether the cost of 
implementing such web vs. the benefit of using it makes it worthwhile.

> 
>>What I am looking for is something similar for this. Another important 
>>point is what John mentioned - ISP MTAs which are sending out spam along 
>>with legit email. For these, unless the percentage of spam becomes 
>>really big, the web of trust will not help with.
> 
> 
>   Why not?  If I know that an ISP sends out 1% spam by accident, I can
> deal with it.  If I know that an ISP sends out 50% spam because
> they're unwilling to solve the problem, I'm better able to deal with
> the traffic.
> 
>   Again, the trust is not, and can not, be a simple binary metric.
> Any measure of trust with is solely "trusted/untrusted" is doomed to
> fail.
> 
>   And the spam problem is not known ISP's sending small volumes.  It's
> substantially unknown people sending large volumes.
> 

I was refering specifically to the use of neutral metrics - there is not 
way to stop spam coming from an ISPs mail server simply by knowing that 
this is a well known MTA. If you use some kind of trust metric, then it 
would be possible actually to do more. Of course that brings us to the 
question of trust.

> 
>>Therefore, IMHO such "web of trust" is more geared towards reducing
>>the problem with hijacked computers, or IPs/domains that have are
>>not supposed to be MTAs. If this in fact is true, than a web of
>>trust method might not be better than other methods to address the
>>problem.
> 
> 
>   You could use a whitelist.  But how would you prime it, if not with
> some kind of web of data?
> 

I will leave it for now, I can't think of other methods offhand. But I 
think that we will revisit this issue in the future.

Yakov