Re: a few short notes

At 23:47 01/02/2004, Jari Arkko wrote:

>Martin Duerst wrote:
>>At 16:56 04/02/01 +0100, Iljitsch van Beijnum wrote:
>>
>>>Mandatory authentication is also a bad idea IMO. Obviously 
>>>authentication is very important and must be supported so that people 
>>>who only want to receive mail from verifyable sources get to implement 
>>>this policy, but that doesn't mean that we should force *everyone* to 
>>>use such a policy.
>>
>>I tend to disagree. *I* don't want you to receive a message that is faked
>>and pretends to come from me, but doesn't. Even if you don't care, I do.
>>I would strongly prefer a system that would not allow this. And my gut
>>feeling is that most users would be with me on this.
>
>I think the hard question is what level of "authentication" to require.
>I am pretty sure 99.99% of users want better protection against spoofed
>addresses than we currently have. But I am not sure they are willing to
>go as far as mandating a global PKI of all e-mail users. So what's left, then?
>Just verifying the two domains but not the users?

I think the 'chain of trust' idea comes in. I want to know that the mail 
server that sent me a message from '[email protected]' actually was 
allowed to do that. Then, I want to know that that mail server had checked 
who it had got the message from (ideally it has put details of the check it 
did into the message's metadata)

Yes, that mail server could fake the authentication details it had 
received, but it could be made so it wouldn't be able to fake whether or 
not it was allowed to send mail from that address. So, if it was allowed to 
send mail from [email protected], but it had faked the authentication 
details it had received, I would have some confidence in blocking the mail 
from that domain, because I know at least that domain hasn't been faked.

The only way for me to know that the original sender IS who (s)he says it 
is, is to have a global PKI with a central registry who verifies all users 
are who they say they are - which isn't possible.

Having the 'chain of trust' idea is a vast improvement over what we have 
now, wouldn't be too hard to implement (IMHO) and would make it a lot more 
difficult (if not impossible) for spammers to work in the way they do now.

"All" it would leave would be ISPs who allow quick & easy (& free, 
sometimes) sign up without checking their new users' details, and then 
allow them to send unlimited emails before their accounts get blocked (if 
the ISP goes that far). I don't see any workable way to stop that 
particular problem other than by those ISPs acting with more care themselves.

Even that would be easier to work around that it is now as anti-spam 
databases could be set up to quickly react to spammer email addresses in 
the same way that DNSBLs are set up now for spammer IP addresses, but with 
more precision, and more accuracy, since you could *know* that the spammer 
email address is authentic.


Paul				VPOP3 - Internet Email Server/Gateway
[email protected]			http://www.pscs.co.uk/