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

Thu Mar 27 19:14:23 UTC 2014

On Mar 27, 2014, at 11:15 AM, Barry Shein <bzs at world.std.com> wrote:

> 
> On March 26, 2014 at 22:25 owen at delong.com (Owen DeLong) wrote:
>> 
>> Actually, a variant on that that might be acceptable… Make e-postage a deposit-based thing. If the recipient has previously white-listed you or marks your particular message as “desired”, then you get your postage back. If not, then your postage is put into the recipients e-postage account to offset the cost of their emails.
>> 
>> Thoughts?
> 
> It's a fine idea but too complicated.
> 
> Look, the (paper) post office doesn't say "oh, you WANTED that mail,
> ok, then we'll return the cost of postage to the sender!"
> 
> Why? Because if they did that people would game the system, THEY'D
> SPAM!

How would they benefit from that?

SPAM — Pay, say $0.10/message.
Then Claim you wanted the SPAM, get your $0.10/message back for each SPAM you sent to yourself.
Or, claim you didn’t want the SPAM and get $0.05/message for each message you received while the
original provider keeps the other $0.05.

> And it would take way too much bookkeeping and fraud identification etc.

Please explain in detail where the fraud potential comes in.

By my interpretation, you’d have to somehow get more back than you deposited (not really possible) in order to profit from sending SPAM this way.

> Let's take a deep breath and re-examine the assumptions:
> 
> Full scale spammers send on the order of one billion msgs per day.
> 
> Which means if I gave your account 1M free msgs/day and could
> reasonably assure that you can't set up 1,000 such accts then you
> could not operate as a spammer.

Not sure how you enforce these user account requirements or how you avoid duplicative accounts.

> Who can't operate with 1M msgs/day?
> 
> Well, maybe Amazon or similar.
> 
> But as I said earlier MAYBE THEY SHOULD PAY ALSO!

I, for one, don’t want my Amazon prices increased by a pseudo-tax on the fact that they do a large volume of email communications with their customers. They have enough problems trying to get IPv6 deployed without adding this to their list of problems.

> We really need to get over the moral component of spam content (and
> senders' intentions) and see it for what it is: A free ride anyone
> would take if available.

I disagree. I see it as a form of theft of service that only immoral thieves would take if available.

> Ok, a million free per acct might be too high but whatever, we can all
> go into committee and do studies and determine what the right number
> should be.
> 
> I'd tend towards some sort of sliding scale myself, 100K/day free,
> 1M/day for $1, 10M/day for $100, 100M/day for $10K, etc. Something like
> that.
> 
> Why would it work?
> 
> Because that's how human society works.
> 
> People who are willing to pay their $10K/mo will demand something be
> done about freeloaders, enforcement has to be part of the cost
> overhead.

But who charges these fees and how do they enforce those charges against miscreants that are sending from stolen hosts, bots, fraudulent IP addresses, etc.?

> And it'd create an economy for hunting down miscreants.

So you’ve got a set of thieves who are stealing services to send vast volumes of email and you want to solve that problem by charging them more for those services that they are stealing (and, by the way, also charging some legitimate users as well).

My guess is that the spammers are going to keep stealing and the people now being taxed for something that used to be free are going to object.

> P.S. And in my vision accepting only email with valid e-postage would
> be voluntary though I suppose that might be "voluntary" at the
> provider level. For example someone like gmail at some point (of
> successful implementation of this scheme) might decide to just block
> invalid e-postage because hey your gmail acct is free! Let someone
> else sell you rules you prefer like controlling acceptance of invalid
> e-postage yourself.

Well, here we get a hint at how you envision this working. There are lots of details that need to be solved in the implementation of such a scheme and I think the devil is prevalent among them.

Owen