why IPv6 isn't ready for prime time, SMTP edition

Robert Drake rdrake at direcpath.com
Sun Mar 30 23:58:37 UTC 2014

On 3/30/2014 12:11 AM, Barry Shein wrote:
> I don't know what "WKBI" means and google turns up nothing. I'll guess
> "Well Known Bad Idea"?
> Since I said that I found the idea described above uninteresting I
> wonder what is a "WKBI" from 1997? The idea I rejected?
> Also, I remember ideas being shot down on the ASRG (Anti-Spam Research
> Group) list primarily because they would take ten years to gain
> acceptance.
> Over ten years ago.
> Maybe they were bad ideas for other reasons. Some certainly were.
> But there's this tone of off-the-cuff dismissal, oh that would take
> TEN YEARS to gain traction, or that's a WKBI, which I don't find
> convincing.
> I read your paper, for example, and said it's a nice paper.
> But I don't find it compelling to the degree you seem to want it to be
> because it mostly makes a bunch of assumptions about how an e-postage
> system would work and proceeds to argue that the particular model you
> describe (and some variants) creates impossible or impractical
> hurdles.
> But what if it worked differently?
> At some point you're just reacting to the term "e-postage" and
> whatever it happens to mean to you, right?
Imagine living in a world where this system is implemented.  Then 
imagine ways to break it.   The first thing I can think of is money 
laundering through hundreds of source and destination email accounts.  
The second is stolen identities or credit cards where the money doesn't 
exist to begin with (Who pays when this happens?)

Third is administrative overhead.  Banks/paypal/exchanges/someone is 
going to want a cut for each transaction, and they deserve one since 
they're going to end up tracking all of them and need to be able to 
reverse charges when something goes wrong.  But then you have a central 
point of failure and central monitoring point so you want to involve 
multiple exchanges, banks, etc.

Then you've got a dictatorship somewhere who says they want an extra 
$0.03 tacked on to each transaction, only it's not $0.03 it's <insert 
famously unstable currency here> so any mail that goes to that country 
has to have custom rules that fluctuate multiple times a day.

Then there is my mom, who knows just enough about computers to send cat 
pictures and forward me chain letters.  She'll not understand that email 
costs something now, or how to re-up her email account when it runs 
out.  The administrative burden will either fall to me or her ISP, and 
each phone call to the ISP probably costs them $$ because they must pay 
a live human to walk someone through email.

> You can't really say you've exhaustively worked out every possibility
> which might be labelled "e-postage". Only a particular interpretation,
> a fairly specific model, or a few.
> When people talked of "virtual currency" over the years, often arguing
> that it's too hard a problem, how many described bitcoin with its
> cryptographic mining etc?
> Bitcoin might well be a lousy solution. But there it is nonetheless,
> and despite the pile of papers which argued that this sort of thing
> was impossible or nearly so.
> Note: Yes, I can also argue that Bitcoin is not truly a virtual
> currency.
> Sometimes a problem is like the Gordian Knot of ancient lore which no
> one could untie. And then Alexander The Great swung his sword and the
> crowds cried "cheat!" but he then became King of Asia just as
> prophesized.
>   >
>   > Regards,
>   > John Levine, johnl at iecc.com, Primary Perpetrator of "The Internet for Dummies",
>   > Please consider the environment before reading this e-mail. http://jl.ly
The answer is that you can't do this to SMTP.  Nobody will ever have the 
answers to all the questions involved with adding cost transactions to 
the protocol.  The only way to do this is to reboot with a new protocol 
that people start to adopt, and the only way they'll do that is if it's 
markedly better than the old way.  You have to remember some people when 
given the choice of paying for email or accepting 10 spams/day will opt 
for accepting a little spam.

The good news is, with email consolidated into 5 or so large providers 
and most people using webmail or exchange, you've got an opportunity to 
change the backend.  Not much software has to be modified, but you do 
need those large providers to buy-in to the idea.

More information about the NANOG mailing list