why IPv6 isn't ready for prime time, SMTP edition
lowen at pari.edu
Wed Mar 26 17:26:35 UTC 2014
On 03/26/2014 12:59 PM, John Levine wrote:
>> That way? Make e-mail cost; have e-postage.
> Gee, I wondered how long it would take for this famous bad idea to
> I wrote a white paper ten years ago explaining why e-postage is a
> bad idea, and there is no way to make it work. Nothing of any
> importance has changed since then.
And I remember reading this ten years ago.
And I also remember thinking at the time that you missed one very
important angle, and that is that the typical ISP has the technical
capability to bill based on volume of traffic already, and could easily
bill per-byte for any traffic with 'e-mail properties' like being on
certain ports or having certain characteristics. Yeah, I'm well aware
of the technical issues with that; I never said it was a good idea, but
what is the alternative?
I agree (and agreed ten years ago) with your assessment that the
technical hurdles are large, but I disagree that they're completely
insurmountable. At some point somebody is going to have to make an
outgoing connection on port 25, and that would be the point of billing
for the originating host. I don't like it, and I don't think it's a
good idea, but the fact of the matter is that as long as spam is
profitable there is going to be spam. Every technical anti-spam
technique yet developed has a corresponding anti-anti-spam technique
(bayesian filters? easy-peasy, just load Hamlet or the Z80 programmer's
manual or somesuch as invisible text inside your e-mail, something I've
seen in the past week (yes, I got a copy of the text for Zilog's Z80
manual inside spam this past week!). DNS BL's got you stopped? easy
peasy, do a bit of address hopping.) The only way to finally and fully
stop spam is financial motivation, there is no 'final' technical
solution to spam; I and all my users wish there were.
More information about the NANOG