FTC Wants Comments on Email Authentication

An anonymous reader writes "Groklaw has the scoop. The Federal Trade Commission and National Institute of Standards and Technology (NIST) will co-host a two-day 'summit' November 9-10 to explore the development and deployment of technology that could reduce spam. The E-mail Authentication Summit will focus on challenges in the development, testing, evaluation, and deployment of domain-level authentication systems. The FTC will be accepting public comments until Sept. 30, 2004 via snail-mail or email (authenticationsummit at ftc.gov). The FTC has a list of 30 questions they would like answers/comments to. The list available in this PDF of the Federal Register Notice." In a related subject, reader Fortunato_NC submits this writeup of the sequence of events that led to Sender-ID's abandonment.

Comments

The Hardest Issue

[Nos.]

Is to keep email easy to use. SPF is a nice idea, but doesn't cope with a couple issues. The first is that a lot of SPAM comes from trojan'd machines. SPF won't prevent or help mark email coming from these machines as SPAM. Secondly, its not expensive to register a domain and flood SPAM for a few days until that domain is blacklisted. Wash, rinse, repeat. I'm not saying a solution isn't out there, just nothing that I have seen really talks to these two issues.

Why not go after the merchants?

[14erCleaner]

You know, I can't figure out why we can't combat spam by making it illegal to send unsolicited ads via email (or maybe the can-spam act already does this), but then go after the companies who are actually trying to get customers. After all, they either provide valid contact information, or nobody can buy from them. If nobody can sell anything via spam any more, the reason for it would go away.

A stopgap measure

[grasshoppa]

An effective stop gap measure would be for ISPs to block port 25 ( along with a number of others ) outbound by default, and open it up only on customer requests.

This way, zombie'd machines wouldn't have a chance to spew their virus/spam emails to everyone, I could still run my home email server, and the ISPs would save on bandwidth.

I wonder why this ISN'T yet in place, to be honest.

Here's the system...

[RecycledElectrons]

Every eMail that is sent (by SMTP - the Simple Mail Transport Protocol) should be considered "unconfirmed." This means that it may or may not be from the return address.

I propose that we add a new layer called CMTP - the Complex Mail Transport Protocol.

CMTP simply takes an unconfirmed eMail (sent by SMTP) and sends a packet back to the sender. This packet asks for verification of the message. The packet includes a checksum, the length, to, from, subject, and the time/date that the eMail was sent.

The sending mail server receives this CMTP checks all of that information, and replies with a CTMP confirmed message or a CMTP not confirmed message.

There is no limit on the number of times that a mail server may be asked to confirm an eMail. There is a limit that messages should not be confirmed more than 24 hours after they are sent. This may pose a small problem in that SMTP does not place a time limit on mail messages.

CMTP does require that every mail server maintain a list of the eMail it has sent. That COULD be time consuming.

CMTP also adds 2 packets to every eMail sent. SMTP was designed to be dead simple. They thought that they could not afford 2 extra packets. In that time, eMail was 80% of all internet traffic. Today, eMail is such a small percentage of all traffic that trpilling it would not be noticed.

Andy Out!

As if you didn't already know this was important..

[museumpeace]

Let me undescore the impact the conference is likely to have by pointing out that when NIST speaks, the DOJ listens. Here is a quote from a rejected submission of mine that found other documents NIST has authored that Ashcroft and co. now use.

Feeding the fascination many /. readers may have for the escalation of technique and counter-technique beteween hackers and computer forensics experts may not be as valuable as keeping clues about how to avoid getting caught out of the hands of the hackers but I just can't resist... Sciencedaily.com pointed me to something hackers and other criminals might want to study carefully: the PDF guidebook that NIST wrote for the DOJ's first responders to computer crime scenes. Though it has John Ashcroft's name at the top, a glance at the document's time line shows that it was authored by experts mostly from outside the DOJ and completed before the current administration's appointments: the imprimatur of Justice Department on the document may not be ironic.

Drat! I'm gonna get modded for flamebait but with a sig like mine, who'd notice?

No mention of sender pays

[gr8_phk]

There was no mention of sender pays postage as a solution. Anything that prevents anonymous email has an inherent central control which the internet doesn't need more of.

Re:NOTHING but an open standard.

[JabberWokky]

Spam is "horrific" and all (BTW I don't get more than 5 a year)

And I get 1800 a day. That's because I am the public contact for several companies with some of my email addresses dating back over 10 years. In conjunction with theater groups and businesses, my email appears in press releases, on fliers, ancient usenet posts, and otherwise all over the place.

Individuals using their email account to talk to friends don't have as much a problem as people who use their email address publically for business and publicity.

My phone number and address are also published. I don't, however, get 1,800 unsolicited calls every day and my junk physical mail is quite reasonable.

--
Evan "I'm not even saying Spam is bad, I'm just saying it costs me serious time"

Re:No Free Software radicals allowed

[JamesTRexx]

Seeing that about 75% of mail is handled by open source mta's, they can't afford to go with ip, moneygrabbing, patentfilled solutions.
The only standard that will get accepted will be an open, patentfree one supported by the free software community.
Any closed or patented ones could only be used between the commercial mta's, so it would have little effect on the amount of spam.

Re:Publish SPF now, be the 126519th...

[wayne]

Actually, I have a list of around 650,000 domains in .COM, .NET and .ORG that have SPF records. These should show up in the SPF Adoption Roll Real Soon Now. Surveys of the .DE and .FR TLDs have also been done, but I don't have the results of those.

I'd like to know how many of those domaines actually are applying effective policies.

In the survey of the .COM domains, I found the top ten SPF records to be:

159416 "v=spf1 mx -all"
147883 "v=spf1 -all"
51245 "v=spf1 ip4:10.0.0.0/24 ip4:10.0.0.0/24 ?all"
28206 "v=spf1 a:smtp.example.net -all"
21437 "v=spf1 mx ip4:10.0.0.0/19 ~all" ""
19733 "v=spf1 mx ~all"
15245 "v=spf1 a:smtp.example.com ~all"
9488 "v=spf1 ip4:10.0.0.0/24 mx -all"
6371 "v=spf1 ip:10.0.0.0/24 ip:10.0.0.0/27 ip:10.0.0.0/24 ip:10.0.0.0/27 ip:10.0.0.0/27 ip:10.0.0.0/27 ip:10.0.0.0/27 ip:10.0.0.0/27 ?all"
5842 "v=spf1 ip4:10.0.0.0/24 -all"
(I have munged the domain names and IP addresses for privacy reasons.)

As you can see, it is very common to define strict SPF record with the "-all" at the end. Those domains that use the softfail option of "~all" are somewhat more lax, but still moving in the right direction.

The complete survey results are available to people who follow the IETF MARID list and/or the SPF discuss list. I'm not going to post a link to them here 'cause I don't want to be slashdotted.