Jacob Palme | 14 Jan 10:27 2001
Picon
Picon

Allowing MTAs to split messages to different recipients

I am working on a compendium for a course I give at my university.
A controversial text from the compendium is quoted below. If you
think that this is wrong, tell me.

--- --- start quote --- ---
If a sender in Europe sends a message to two or more recipients in North America, only one copy might be copied across the expensive Atlantic cables as shown in the figure below:


A problem with this, however, is that most MTAs are not willing to
handle mail, unless either the recipient or the sender is local to
the MTA. Thus, the saving shown above requires an agreement with the
MTA which splits the message after transport across the Atlantic.
This was not always so. In the beginning of the 1990-s, most MTAs
were willing to forward mail for any recipient. The reason why this
was abolished in the middle of the 1990-s was that spammers used this
feature to get foreign MTAs to help them split mail to millions of
recipients. Some so-called experts claimed that spamming could be
stopped by forbidding splitting of mail by other than the MTA of the
sender or the recipient. They enforced their view by implementing a
program which scanned all MTAs everywhere, checking that they did not
allow foreign splitting, and sending angry letters to non-conforming
MTA administrators (postmasters) threatening to stop receiving mail
from them unless they stopped splitting. This is an interesting
example of how the Internet is regulated in dubious ways by
pseudo-police-authorities. Spamming could be counteracted more
efficiently using other methods than this.
--
Jacob Palme <jpalme <at> dsv.su.se> (Stockholm University and KTH)
for more info see URL: http://www.dsv.su.se/jpalme/
chicks | 15 Jan 16:42 2001
Picon

Re: Allowing MTAs to split messages to different recipients

This is from a text?  It sounds more like a spammers diatribe.  The spin
to the whole piece is way overdone even if you disagree with what various
people did.  Mentioning the word "relay" and defining it would seem to be
a good idea.  Providing some links to the well-reasoned opposing opinions
would help add some balance.  But the tone of the whole thing is way off.

--

-- 
</chris>

Those who cannot remember the past are doomed to buy Microsoft products.

Keith Moore | 15 Jan 17:19 2001
Picon

Re: Allowing MTAs to split messages to different recipients

> If a sender in Europe sends a message to two or more recipients in 
> North America, only one copy might be copied across the expensive 
> Atlantic cables as shown in the figure below:
> 
> A problem with this, however, is that most MTAs are not willing to
> handle mail, unless either the recipient or the sender is local to
> the MTA. 

There are more fundamental problems than this, namely, that the
sending MTA has no idea where its recipients are located with
respect to expensive links, so it has no basis with which to optimize. 

> Thus, the saving shown above requires an agreement with the
> MTA which splits the message after transport across the Atlantic.
> This was not always so. 

True, but probably not relevant.

> In the beginning of the 1990-s, most MTAs
> were willing to forward mail for any recipient. The reason why this
> was abolished in the middle of the 1990-s was that spammers used this
> feature to get foreign MTAs to help them split mail to millions of
> recipients. Some so-called experts claimed that spamming could be
> stopped by forbidding splitting of mail by other than the MTA of the
> sender or the recipient. They enforced their view by implementing a
> program which scanned all MTAs everywhere, checking that they did not
> allow foreign splitting, and sending angry letters to non-conforming
> MTA administrators (postmasters) threatening to stop receiving mail
> from them unless they stopped splitting. This is an interesting
> example of how the Internet is regulated in dubious ways by
> pseudo-police-authorities. Spamming could be counteracted more
> efficiently using other methods than this.

While your history is reasonably accurate, I think you're mixing
two fairly unrelated things. And I don't think you've supported 
your last statement.  

In general it doesn't seem reasonable to expect an MTA to handle 
inbound traffic for other recipients than those for whom it's agreed 
to exchange mail.  Regardless of whether or not the traffic is spam, 
it's still taking up their resources.  To expect a remote MTA to
forward third party traffic, in the absence of explicit agreement
and permission to do so, strikes me as naive.  Even though at one
time it was a common practice, it's never been required by the 
SMTP protocol, and it was therefore never safe from the point-of-view
of application design to depend on this happening.  On the other
hand, if you get explicit agreement from remote MTAs to forward
your list mail (with either authentication or narrowly tailored
filtering for your list traffic), you can still do this.

To my way of thinking, the blacklists were indeed abusive.  They 
were mounting a distributed denial-of-service attack, using
propaganda and disinformation to entice gullible MTA administrators
into filtering using their criteria.  To me this is little better
than putting a virus in an executable attachment of an email message 
and labelling it in such a way that the recipient is likely to click
on it.  Perhaps they had good intentions, but some virus writers have
good intentions also.  In this case the attackers were naive 
about the likely good that it would do and about the harm that would
result.  But from my point-of-view the principal damage that was done was 
to thwart the legitimate use of SMTP for submission of messages by mobile 
users to their "home" SMTP servers.  (and no, SMTP-after-POP doesn't 
work reliably, and SMTP AUTH isn't widely enough deployed even now)

Keith

ned.freed | 15 Jan 16:15 2001

Re: Allowing MTAs to split messages to different recipients

> I am working on a compendium for a course I give at my university.
> A controversial text from the compendium is quoted below. If you
> think that this is wrong, tell me.

OK, I'll tell you: It's wrong on many points.

> If a sender in Europe sends a message to two or more recipients in
> North America, only one copy might be copied across the expensive
> Atlantic cables as shown in the figure below:

> A problem with this, however, is that most MTAs are not willing to
> handle mail, unless either the recipient or the sender is local to
> the MTA.

Or unless the sender has appropriate credentials. These days locality isn't
sufficient, since many users roam and still want to be able to use the MTA at
their home site as a relay.

> Thus, the saving shown above requires an agreement with the
> MTA which splits the message after transport across the Atlantic.

Yes, but such agreements are easily accomodated by present-day infrastructure
(SASL and TLS). Nevertheless, AFAIK there are relatively few such setups.
Possible reasons why this is so include:

(1) This practice, while wasteful, is small beer compared to other uses of
    these links.

(2) The people for whom it is a problem aren't the same as the people who
    are able to set up such arrangements.

(3) People are ignorant of just how easy this problem is to solve. (It would
    only take a few minutes to configure the software I work on to support
    this sort of relay, for example, and I have no reason to believe it is any
    harder in other software.)

> This was not always so. In the beginning of the 1990-s, most MTAs
> were willing to forward mail for any recipient. The reason why this
> was abolished in the middle of the 1990-s was that spammers used this
> feature to get foreign MTAs to help them split mail to millions of
> recipients.

Yes, and this is why the relaying got shut down and continues to be shut down.
Try digging yourself out of having your system used for a massive spam relay
sometime -- once you've done so cries about how this leads to such inefficiency
on high-cost links somehow fail to impress.

> Some so-called experts claimed that spamming could be
> stopped by forbidding splitting of mail by other than the MTA of the
> sender or the recipient. They enforced their view by implementing a
> program which scanned all MTAs everywhere, checking that they did not
> allow foreign splitting, and sending angry letters to non-conforming
> MTA administrators (postmasters) threatening to stop receiving mail
> from them unless they stopped splitting. This is an interesting
> example of how the Internet is regulated in dubious ways by
> pseudo-police-authorities.

While I have little love for the scan-and-ban crowd, you're giving them too
much credit here.

The reason open relays were shut down is because a significant number of them
were exploited to the point where the users of those systems were unable to get
useful work done. Most administrators have no choice but to block open relay,
regardless of its utility. And this situation hasn't changed -- try opening a
system whose name is widely known up to relay sometime and you'll see.

The only reason that the scan-and-ban crowd has been at all successful is
because most systems already blocked open relay and they were able to go after
the few remaining ones that didn't. Had the situation been one where the common
practice was to allow open relay they would have gotten nowhere: Subscribing to
their "don't accept mail from systems that permit open relay" policy would have
resulted in too much mail being blocked to be useful.

Indeed, it isn't entirely clear to me that the real weapon in the scan-and-ban
crowd's arsenal is the number of sites subscribing to their black list. I've
seen several cases where a site that was open to relay was blacklisted and then
almost immediately was used by a spammer for the first time. It is undeniable
that the list of open relays these folks compile could be quite useful to
spammers, and while I presume they take steps to try and prevent their data
from being used by spammers, I find the timing above to be more than a little
suspicious.

> Spamming could be counteracted more
> efficiently using other methods than this.

And these methods are what exactly? If there's something more effective than
open relay blocking I'd like to hear what it is.

				Ned

D. J. Bernstein | 15 Jan 22:24 2001
Picon

Re: Allowing MTAs to split messages to different recipients

Jacob Palme writes:
> The reason why this was abolished in the middle of the 1990-s was that
> spammers used this feature to get foreign MTAs to help them split mail
> to millions of recipients.

You have the timeline wrong. As far as I know, there was only one MTA
that turned off relaying by default before 1997. Sendmail didn't do that
until 1998. Open relays were still widespread in 1999.

---Dan

Harald Alvestrand | 16 Jan 00:37 2001
Picon

Re: Allowing MTAs to split messages to different recipients

At 10:27 14/01/2001 +0100, Jacob Palme wrote:
>I am working on a compendium for a course I give at my university.
>A controversial text from the compendium is quoted below. If you
>think that this is wrong, tell me.

If it is a course in the politics and psychology of self-regulation of the 
Internet, I think it entirely appropriate, if labelled with "IN THE OPINION 
OF THIS PROFESSOR", and with appropriate representation of alternate views 
given.
As formulated, none of these criteria are fulfilled.

If it is a course in the mechanics of email propagation, I find it 
inaccurate, inappropriate and irrelevant.

It all depends on the context.
--
Harald Tveit Alvestrand, alvestrand <at> cisco.com
+47 41 44 29 94
Personal email: Harald <at> Alvestrand.no

Jacob Palme | 18 Jan 02:35 2001
Picon
Picon

Re: Allowing MTAs to split messages to different recipients

At 11.19 -0500 01-01-15, Keith Moore wrote:
>But from my point-of-view the principal damage that was done was
>to thwart the legitimate use of SMTP for submission of messages by mobile
>users to their "home" SMTP servers.  (and no, SMTP-after-POP doesn't
>work reliably, and SMTP AUTH isn't widely enough deployed even now)

It is true that this is the main reason why I get so angry with
this. Having to find the right acceptable SMTP relayer wherever
I go is a pain. And having to send my messages one at a time
directly to the recipient local MTA is a pain with Eudora,
because Eudora is not able to this automatically. I do not
know if other mail clients can do this. Earlier this week,
I was in Bologna and sent my messages in this way, changing
the receiving SMTP server manually between each message to
the server indicated by the MX record of the recipient. It
worked, but was not easy.

I guess what I should get is some simple MTA which I can
run on my personal computer and which does this for me.
It need not handle incoming mail, only outgoing mail.

SMTP AUTH or SMTP-after-POP seems like a possible solution to
the needs of travelling mail users. Are there standards on
this? Why have they not been accepted? Do they use simple
passwords, or do they use crypthographic authorisation -
such authorisation seems to have much difficulty getting
accepted.

--

-- 
Jacob Palme <jpalme <at> dsv.su.se> (Stockholm University and KTH)
for more info see URL: http://www.dsv.su.se/jpalme/

Jacob Palme | 18 Jan 02:33 2001
Picon
Picon

Re: Allowing MTAs to split messages to different recipients

At 08.15 -0700 01-01-15, ned.freed <at> innosoft.com wrote:
>  > I am working on a compendium for a course I give at my university.
>>  A controversial text from the compendium is quoted below. If you
>>  think that this is wrong, tell me.
>
>OK, I'll tell you: It's wrong on many points.

I sort of expected this reply, that is why I wrote my message
to get more opinions on this issue. I will rewrite the text
based on the valuable input you and others have given on
my flame.

At 08.15 -0700 01-01-15, ned.freed <at> innosoft.com wrote:
>  > Spamming could be counteracted more
>>  efficiently using other methods than this.
>
>And these methods are what exactly? If there's something more effective than
>open relay blocking I'd like to hear what it is.

Relay blocking certainly is not effective. Spamming just increases
every year. Of course, you might claim that without relay
blocking, spamming would have increased even more.

Relay blocking was in reality not meant to counteract spamming,
but to reduce the load of spammers taking over MTAs to forward
their spams. Thus, it was not meant to make life more bearable
to users, but to make life more bearable to MTA owners. But
those who introduced relay blocking wrongly claimed that it stop
spamming.

If relay blocking had been effective, it would be equally effective
if it did not stop smaller amounts of relaying, but only massive
amounts of relaying.
--

-- 
Jacob Palme <jpalme <at> dsv.su.se> (Stockholm University and KTH)
for more info see URL: http://www.dsv.su.se/jpalme/

Keith Moore | 18 Jan 04:22 2001
Picon

Re: Allowing MTAs to split messages to different recipients

> Relay blocking certainly is not effective. Spamming just increases
> every year.

relay blocking doesn't stop spam from getting to recipients.
it does, however, lessen the load on MTAs that were being abused by spammers.

Keith

Harald Alvestrand | 18 Jan 11:43 2001
Picon

Re: Allowing MTAs to split messages to different recipients

At 02:33 18/01/2001 +0100, Jacob Palme wrote:
>Relay blocking certainly is not effective. Spamming just increases
>every year. Of course, you might claim that without relay
>blocking, spamming would have increased even more.
>
>Relay blocking was in reality not meant to counteract spamming,
>but to reduce the load of spammers taking over MTAs to forward
>their spams. Thus, it was not meant to make life more bearable
>to users, but to make life more bearable to MTA owners. But
>those who introduced relay blocking wrongly claimed that it stop
>spamming.

Relay blocking is also effective against trace hiding.
If the spammer has to send mail from his own account, finding it becomes 
easier.

--
Harald Tveit Alvestrand, alvestrand <at> cisco.com
+47 41 44 29 94
Personal email: Harald <at> Alvestrand.no


Gmane