Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Ok Google: please publish your DKIM secret keys

blog.cryptographyengineering.com

Post reply on HN
Print view View original

231–240 of 492 posts

Re: Ok Google: please publish your DKIM secret keys

#231

Post by thinkharderdev »

quit32 wrote: post #194
avianlyric wrote: post #120

Earlier quoted context omitted.

If non-repudiation is important to you, then both parties should consent to it and use a platform that explicitly supports it. It shouldn’t be sprung on people without consent. It would be like saying it’s fine to keep a recording from someone else’s webcam because it might prove a crime later. There’s a reason why justice systems have statues of limitations. People should need to look over their shoulders for the re…

It is not really being sprung on them with how long it has existed. Not at all like continuing recording on a webcam where you might say things never intended for the party receiving it. Are ppl who don't even know DKIM exists but know they have shady emails saved in the cloud or on their personal really just banking on repudiation and thats why they take no other action like deleting the email or putting more though…

I think it is less about user behavior and understanding than it is about the incentives it creates for nefarious actors. Basically, if emails can't be cryptographically verified then stealing a bunch of emails and anonymously dumping them somewhere is pointless since most people would probably not consider them authentic.

Re: Ok Google: please publish your DKIM secret keys

#232

Post by baobabKoodaa »

hyperpape wrote: post #198
baobabKoodaa wrote: post #75

Earlier quoted context omitted.

> once counterparties have authenticated each other's messages, the legitimate need for authentication is gone; allowing random strangers to authenticate messages concedes information to them. As you know, there are many legitimate needs to authenticate messages of strangers. For example, when you order products over the internet, an e-mail of your purchase is often the only proof of what was agreed in the purchase.…

As you know, there are many legitimate needs to publish naked pictures of your body to the entire internet, but most people want to keep them private. Defaults matter. What makes this hard is that email is responsible for too much crap. No single user interface should carry: 1. Party invites 2. Private messages to your spouse, therapist, pastor, etc. 3. Marketing messages 4. Password recovery requests 5. Financial tr…

I agree that e-mail is horrible, and I would prefer if contracts (like purchase receipts) moved off e-mail to something else, but that's tangential to the point. If Google started to publish and rotate DKIM keys, the world wouldn't suddenly move away from email.

Re: Ok Google: please publish your DKIM secret keys

#233

Post by microtherion »

Alex3917 wrote: post #56
xbar wrote: post #25

I think this is a shameful argument. Non-repudiation over time is a truly powerful property of DKIM'd email for a great many uses outside of blackmail. Calling for the ability to remove it during the years 2016-2020 in order to "protect politicians from blackmail" is not only of deeply questionable value but of suspect motivation. Who is the author interested in protecting?

> Non-repudiation over time is a truly powerful property of DKIM'd email for a great many uses outside of blackmail. This. Publishing the DKIM keys would be a huge loss for email archivists and historians in general. E.g. a couple weeks ago Donald Knuth published all of the emails he's sent and received over the last 20+ years of his career[1], without DKIM how would we know that they are authentic? [1] https://libra…

That just proves that he did send the mails he published. But since he did not put the mails on a block chain, we'll never know whether he was e.g. exchanging steamy e-mails with Grace Hopper or arranging weed deals with E.W. Dijkstra (or vice versa) on the side and decided to omit them from his published correspondence >:-)

Re: Ok Google: please publish your DKIM secret keys

#234

Post by lxgr »

baobabKoodaa wrote: post #75
tptacek wrote: post #55

Earlier quoted context omitted.

Among messaging cryptographers, it's not even an argument. Serious secure messengers have been designed to avoid non-repudiation since OTR. Non-repudiation is a vulnerability: once counterparties have authenticated each other's messages, the legitimate need for authentication is gone; allowing random strangers to authenticate messages concedes information to them. Here, have a link, from 2004: https://otr.cypherpunks…

> once counterparties have authenticated each other's messages, the legitimate need for authentication is gone; allowing random strangers to authenticate messages concedes information to them. As you know, there are many legitimate needs to authenticate messages of strangers. For example, when you order products over the internet, an e-mail of your purchase is often the only proof of what was agreed in the purchase.…

> As you know, there are many legitimate needs to authenticate messages of strangers.

Absolutely, but this should be an opt-in feature (and not provided server-side, at that).

Re: Ok Google: please publish your DKIM secret keys

#235

Post by kennywinker »

advisedwang wrote: post #12

If Google did rotate the keys, it's quite likely some enterprising people would log observed keys. With access to a log of keys (that you don't believe has been tampered with) then rotating keys doesn't prevent non-repudation. An alternative would be for email servers to strip the DKIM headers on inbound emails after recording that the email was validated. The email stored at rest then no longer provides non-repudati…

I'm not a fan of offloading that trust onto my email host. Not that I validate dkim headers, but the fact that I have access to them keeps my email host honest.

Re: Ok Google: please publish your DKIM secret keys

#236

Post by Red_Leaves_Flyy »

lxgr wrote: post #234
baobabKoodaa wrote: post #75

Earlier quoted context omitted.

> once counterparties have authenticated each other's messages, the legitimate need for authentication is gone; allowing random strangers to authenticate messages concedes information to them. As you know, there are many legitimate needs to authenticate messages of strangers. For example, when you order products over the internet, an e-mail of your purchase is often the only proof of what was agreed in the purchase.…

> As you know, there are many legitimate needs to authenticate messages of strangers. Absolutely, but this should be an opt-in feature (and not provided server-side, at that).

>this should be an opt-in feature (and not provided server-side, at that).

Why?

Re: Ok Google: please publish your DKIM secret keys

#237

Post by hamburglar »

adrianmonk wrote: post #206
some_furry wrote: post #33

Earlier quoted context omitted.

> Can't you just set up your mailserver so that it drops all the crypto headers (DKIM-Signature, ...) after verifying them and storing the result in Authentication-Results? Matthew Green's ask isn't about protecting users that are tech-savvy enough to just set up their own mailserver and configure it a special way. It's about protecting the billions of users that aren't.

Gmail (and other email providers) could also protect these billions by making the header-stripping change at the server level for everyone. After all, Green is proposing for them to change their servers anyway, so either way it requires some kind of server change. The advantage of Green's approach is it gets results quickly because with one change they can protect a lot of emails. But, while quick results are nice, i…

Yes, the call isn't for GMail to kill non-repudiation protections on mail they receive (and was signed by others), it's to kill it on emails they've signed and thus are sitting on someone else's server.

Re: Ok Google: please publish your DKIM secret keys

#238

Post by mthoms »

busterarm wrote: post #215
mthoms wrote: post #165

Earlier quoted context omitted.

Using the billionaire pedophile example is a disgusting trick. You're trying to set me up for appearing to support that. Why not use more neutral examples? Like being gay or supporting certain political causes? What if those later become controversial or illegal? What then? Do you want to live in a world where you have to guard everything you say in semi-private conversations, just in case it one day becomes controve…

I'm not setting you up as supporting anything. I'm illustrating the severity of the identifiable public record. When the Nazis started rounding up people to put in camps, they looked at the _extremely detailed_ Christian Parish records saying who was what and where they lived. They were thought to be innocuous and important records to keep at the time. Actually I think in the Scandinavian countries the state Church i…

>I'm illustrating the severity of the identifiable public record.

This is an argument in favor of emails being non-verifiable, so now I'm confused.

Previously you seem to be supporting the idea that email should be verifiable. Now you seem to be arguing the opposite. Everything you wrote above correlates with the opinions I've expressed so far.

You also wrote:

> If you live in a [country where homosexuality is illegal] and secretly a homosexual, do not send emails indicating that you're a homosexual.

As if that's an acceptable state of affairs and a reasonable compromise for the purpose of catching the occasional bad actor. It isn't.

Re: Ok Google: please publish your DKIM secret keys

#239

Post by CodesInChaos »

arkadiyt wrote: post #19
CodesInChaos wrote: post #13

Earlier quoted context omitted.

You can authenticate messages in a way that only the recipient can verify (Diffie-Hellman plus MAC).

If you had access to a public key for every email address then why stop at authentication - you could encrypt all email on the web. But we don't, so we can't.

Authentication in this context doesn't need to be end-to-end. Instead of a custom protocol, we could probably just use SSL client certificates authenticating the sending domain to achieve the same effect.

Re: Ok Google: please publish your DKIM secret keys

#240

Post by baobabKoodaa »

hamburglar wrote: post #221
baobabKoodaa wrote: post #161

Earlier quoted context omitted.

Ok, my GPG example was wrong. And yes, you got me, I'm not a professional cryptographer. But can you address the point? You said "once counterparties have authenticated each other's messages, the legitimate need for authentication is gone". I provided a counter-example to demonstrate that your statement was an exaggeration. You clearly dispute some part of this, but it's unclear to me what the disputed part is. Edit:…

> Ok, my GPG example was wrong. And yes, you got me, I'm not a professional cryptographer. But can you address the point? You said "once counterparties have authenticated each other's messages, the legitimate need for authentication is gone". I provided a counter-example ... I think the person you're talking to thinks this is very obvious and thus isn't stating it explicitly, but in the special case where you want an…

> the sender should just add non-repudiation to it in the form of a signature that's intended for that. Simple.

Ok, but this does not magically happen if Google publishes and rotates their DKIM keys. People will continue to use email for everything, but now emails can no longer be authenticated by third parties.

Post reply on HN