Re: Fee Based Email (From Re: Process's PreciseMail AntiSpam...)

david20_at_alpha1.mdx.ac.uk
Date: 10/01/03 

Date: Wed, 1 Oct 2003 00:24:39 +0000 (UTC)

In article <blbctn$4jd$1@n.ruf.uni-freiburg.de>, gartmann@non.immunbio.mpg.de.sens (Christoph Gartmann) writes:
>In article <blabbl$ivi$1@news.mdx.ac.uk>, david20@alpha2.mdx.ac.uk writes:
>
>Won't be the case. A small ISP is connected to a larger ISP, this one is then
>connected to a nation-wide network provider which in turn is linked to some
>international network provider. The bills always go between directly linked
>partners. They already do business which each other.
>
>>Receiving ISP blocks all mail from hotmail ?
>>
>>There are some ISPs like hotmail which are so widely used that it is
>>politically impossible to block them even if they are harbouring tons of
>>spammers. I'd be lynched by my management if I suggested blocking hotmail.
>
>It is a matter of hotmail's ISP only. This one has to decide what to do if it
>gets bills from his uplink - pass them on to hotmail, increase the monthly rate
>for hotmail or drop the contract.
>
>>Port 25 connections tell you a connection attempt was made. Doesn't tell you a
>>mail message was delivered. If a message has to be retried because of problems
>>with the receiver's system the sender would be being charged for every delivery
>>attempt even if this ran into thousands of delivery attempts.
>
>Once again, it is not the sender that would be charged in my approach. It is
>like the following:
>
> sender --> sender's ISP --> ISP's uplink --> some large carrier -->
> larger carrier --> international net company --> large carrier -->
> smaller carrier --> recipient's ISP --> recipient
>
>In my approach either the "larger carrier" or the "international net company"
>will have to pay a fee for each message they route through. So "international
>net company" will bill "larger carrier". Then it is up to "larger carrier" what
>to do - most likely bill "some large carrier" either on a per mail basis or
>with a higher flat fee or implement a scheme like "1000 mails per day are
>included, 10 US$ for each 100 additional ones" or anything in this direction.
>There is no bill going from "recipient" to "recipient's ISP" or from
>"recipient's ISP" to "smaller carrier".
>
>Now let's assume that "recipient" will send a message to "sender". In this case
>"international net company" (the one that is obliged to the tax) will want to
>bill "large carrier". Again, even today they do business with each other, they
>exchange bills, nothing new is required. "large carrier" has to decide whether
>it will bill "smaller carrier" and if so, in which way. It is then "smaller
>carrier's" job to bill "recipient's ISP" and "recipient's ISP" may decide to
>bill "recipient". It doesn't matter in which country "recipient" is located or
>where "sender" is living and in what currency they pay.
>
>Now to the case where delivery of a mail is unsuccessfull: it doesn't matter if
>it counts as a message. The amount of these cases is neglectable to the overall
>number of successfully delivered messages. On the other hand, it is technically
>possible to do an exact billing, but I think it is not worth the overhead.
>
As a percentage of successfully delivered mail messages that maybe so but to
the user who sent out one mail message and is billed for a thousand it is a
different matter.

>>Also this would be a lovely new game for hackers. Spoof an IP address and
>>connect repeatedly to port 25 on a "fee based systems" mailhub.
>
>This won't happen. Look at my example above: sender will get a heavy bill from
>"sender's ISP". The original bill will come from "international net company"
>and will be directed to "larger carrier". This one will pay and request the
>money from "some large carrier" and so on until "sender's ISP" will want the
>mony from sender. I am quite confident that "sender's ISP" will have the means
>to get what it wants or that "sender's ISP" will long have implemented
>something that prevents "sender" from performing costly things that would ruin
>"sender's ISP". "sender" may spoof it's IP address, nevertheless his packets
>originate from him and "sender's ISP" is in a position to tell where the
>packets came from and whom to bill.
>
>I don't seen any organizational problem with my approach. There are some
>technical problems but they are easy to solve. The SMTP protocol as such is
>simple. At a first glance a scanning for packets destinated to port 25 and
>containing "CR.CR" would be reasonably sufficient.
>

You seem to be assuming that mail and the internet are a hierachical network.
They aren't. SMTP is routed over a virtual connection directly between the
sender and recipient systems (Organisations may force this to also pass through
intermediate systems ie their central mailhubs if they wish).
However the link is a virtual connection. The individual packets can be routed
all around the world.
Many companies have multiple connections onto the internet. Many companies
operate in several countries and as well as having internet connections in each
country also have private network connections which in some cases cross
national borders.
This is what gives the internet it's stability. All packets in a single mail
message won't all go through any particular "international net company" let
alone all mail from or to a particular system.

David Webb
VMS and Unix team leader
CCSS
Middlesex University

>Regards,
> Christoph Gartmann
>
>--
> Max-Planck-Institut fuer Phone : +49-761-5108-464 Fax: -452
> Immunbiologie
> Postfach 1169 Internet: gartmann@immunbio dot mpg dot de
> D-79011 Freiburg, Germany
> http://www.immunbio.mpg.de/home/menue.html

Relevant Pages

  • Re: Queue Drops
    ... ISP uses Frame Relay for their Internet connections. ... I have two T1s in one site and three T1s in another connected to the ISP. ... The three T1s are bundled in a Multilink Frame Relay connection. ...
    (comp.dcom.sys.cisco)
  • Re: Using Two ADSL Internet Connections Simultaneously
    ... > connections to the internet and use these connections to provide ... > internet access to several wireless laptop computers. ... same thing (and I suspect Cisco routers at the ISP can do it). ... with "from" IP addresses assigned to the DSL lines. ...
    (comp.dcom.telecom)
  • Re: Strange netstat output - possible hacking attempt?
    ... >> think we can really call that 'port scanning' in any illegitimate sense. ... > out to the colo swerver, the ISP would cut the link, outgoing packets would ... "Requests per 10 seconds per host rule" and only inforcing these rules ... connections making it a WAN. ...
    (comp.os.linux.security)
  • Re: handling hsrp connections from isp
    ... I've got two connections to the same ISP (connected to two of their ... routers), with HSRP running on their routers. ... Load balancing across connections is not a concern here -- I am just ... But the question is dealing with the two HSRP connections from the ...
    (comp.dcom.sys.cisco)
  • Re: handling hsrp connections from isp
    ... Because you only have unmanaged switches for your ISP and Firewall ... connections, that is definitely a single point of failure. ... pair of switches, which then go to the firewalls, which then go back ...
    (comp.dcom.sys.cisco)