Ramblings of an old Doc

 

MS, Google, PayPal, Facebook, AOL and LinkedIn along with several other net heavyweights and Bank of America as well as Fidelity Investments are getting together to put an end to phishing.

No new technology, just a new framework for email using SPF (Sender Policy Framework) and DKIM (DomainKeys Identified Mail). DMARC.org published its guidelines and specification for this new framework.

“What DMARC adds is a policy-based framework of actions and reporting that email providers will follow to act on instructions from enterprise email managers to identify or even block spoofed mail exploiting any enterprise domain name. "We came together to produce a new standard, not a new technology," says McDowell. "This leverages SPF and DKIM, and it puts an end to spoofing, the most common form of email abuse.

"Our goal with the launch of DMARC is we want people to start using it, and improve their email authentication infrastructure," says Adam Dawes, product manager at Google's mail team. "The most widely used tactic for phishing is domain spoofing. It's extremely easy for phishers to take advantage."He said Google is already blocking fraudulent messages based on cooperation through DMARC with Facebook, LinkedIn and PayPal, for example. He said any mailbox hosted by Google has DMARC capabilities with them. Google itself has implemented DMARC "so we can report fraudulent messages that claim to come from Google.com."

 http://www.computerworld.com/s/article/9223807/Google_Microsoft_Facebook_Bank_of_America_team_to_wipe_out_phishing?taxonomyId=17&pageNumber=1

So far, it’s for large institutions, but ISP’s could deploy this needing a DNS network administrator who knows how to aggregate files to send to provider companies like DMARC who will get the companies to verify the emails are indeed from them.

Thus, phishing through spoofing (the most common crime mode) may be coming to an end.

Source:

http://www.computerworld.com/s/article/9223807/Google_Microsoft_Facebook_Bank_of_America_team_to_wipe_out_phishing?taxonomyId=17&pageNumber=3


Comments
on Jan 30, 2012

I'm a bit paranoid around SOPA times, and this coming from the largest corporate privacy invaders in the world.

Implications for the future? What if they decide company X can't do business due to some reason? Think Piratebay, or Wikileaks. Excluded from such a system, if widespread, could be devastating.

on Jan 30, 2012

Okay, let me tell you about my experience with SPF from the point of a system administrator. 

SPF breaks existing MTA mechanisms like mail forwarding. You set a forward from your Google account to your Yahoo account, and voila, your mails get rejected by SPF, because they come from different IP addresses than those specified in the SPF records for the sender domain. Also, to make SPF work, you would have to persuade every receiving mailserver to discard mails that won't pass SPF validation. How users react? "My mails aren't coming through? What? SPF? What's that? I don't care, turn it off! I need my mails, I am losing money here..." Bye bye SPF. The other "workaround" people use to make sure their mail is always received on SPF-enabled servers is that they publish SPF records that allow everyone to send from their domain: "v=spf1 +all"

The basic rule states that security is always a procedure, not a technology or product or application. Sure, there are more and less secure applications, but how secure they really are depends on how they are used. Moreover, security and user comfort are in inverse proportion - the more secure procedure, the more "hassle" it brings. And people want no hassle, so security goes out of the window. 

If you want secure mail, you can have it. The tools are in place for some time - DNSSec and PGP. How? Just follow these steps:

1) Sign your zone with DNSSec

2) Publish PGP public keys for every person that will be sending from your domain via TXT records in your domain. Make sure the keys are sufficiently strong and that the private keys are handled with care (password protected, preferably on encrypted filesystems)

john_doe    IN       TXT     "<public key>"

3) Sign every mail you send with the keys you published in your DNS zone.

4) Persuade every recipient of your mails to discard any mail that is not correctly signed with the correct keys.

There, problem solved. The problem is of course point 4 - people are either lazy, or ignorant, or both. "I don't have time for this". Or similar excuses. You can either have comfort, or security.

SPF has been pushed by Microsoft for YEARS, with negligible results. 

Do you want to really solve SPAM forever? Get rid of the SMTP protocol. It is very bad for today's use, where internet is a hostile place to be. Replace it with a securely designed protocol that requires a cryptographic authentication on every step, like DNSSec. Or forget it. Ain't gonna happen. Too expensive. Too much hassle.

Oh, and when establishing trust chains, don't trust PKI. PKI is broken:

https://mail1.eff.org/pipermail/observatory/2011-September/000308.html

 

Read about the amazing failure of the Diginotar CA on various articles. They kept issuing certificates months after it was obvious they are compromised. SSL 3.0 is broken too.

Ultimately, phishing and scam success boils down to the intelligence of the user. People don't read the certificates browser presents, they just click "Okay, I trust this, make an exception". They are willing to believe a former Nigerian oil minister wants to give them millions of dollars. As long as they exist, frauds of all kinds will prevail.

 

on Jan 30, 2012

Kamamura_CZ
They are willing to believe a former Nigerian oil minister wants to give them millions of dollars.

You mean they don't? !!!!!!!!!!!!!!!

 

 

I'm shattered.....

on Jan 30, 2012

And here I was believing I had a friend who won the Lottery.