Fallacious use of 'bank' in net payment systems

[Ian Grigg]

someone wrote:
> 
> On Sat, May 17, 2003 at 11:52:20AM -0400, Ian Grigg wrote:
> >   * Each ISP already has a relationship with
> >     each other ISP,
> 
> Wrong.  From the packet-level transport view, an ISP only has a
> relationship with its direct neighbors (its transits, customers, &
> peers).  From the end-to-end tcp view, an ISP may have no control
> or knowledge of either end of the connection.  And even if one end
> of a mail connection is one of the ISP's mail servers, the other
> end can belong to anyone - ISP or other - anywhere on the Internet
> & the ISP may have no relationship with them at all.

One ISP can send packets to another (IP), and one
ISP can forward email to another (SMTP).  That's
the relationship.  As written up by Adam B, either
of these protocols can be extended or piggy backed
to carry non-core anti-spam tokens.

Compare this to banks:  one bank can send
a payment to another bank.  (At a pinch, a
bank can even manage to direct the payment
to the right user.)

But, there is no relationship between any bank
and any ISP.  There are only customer-provider
relationships between pairs of these agents.

In net terminology, relationship might be
rewritten as 'protocol'.  In which case ISPs
talk IP, SMTP, etc.  Banks talk SWIFT, FedWire,
BACS, etc.

> >   * Each ISP already knows who is mailing whom,
> >     so there is no need to preserve the privacy
> >     of payments.
> 
> Wrong.  I handle my own email.  My ISPs (the one that I buy
> packet-level transport from) never see anything about my mail other
> than as bits in packets that I send to them.  My mail never goes
> to my ISP's mail servers (unless I happen to be sending mail to
> someone who is using my ISP's mail server).

As far as email is concerned, you are your
own ISP.

-- 
iang

---------------------------------------------------------------------
The Cryptography Mailing List
Unsubscribe by sending "unsubscribe cryptography" to majordomo at metzdowd.com