Re: [Unbearable] Possible attack on Token Binding with RSA key exchange

[Andrei Popov <[e-mail]>]

> Would it be reasonable to say that defending against attackers with possession of a server's private key is outside of the scope of Token Binding?
Server's private key, and/or client private key, and/or master secret... anything that enables the attacker to MITM a user's TLS session, or to duplicate a user's EKM on the attacker's TLS connection, will break TB.

-----Original Message-----
From: Nick Harper [mailto:[e-mail]] 
Sent: Tuesday, September 5, 2017 5:23 PM
To: Andrei Popov <[e-mail]>
Cc: Subodh Iyengar <[e-mail]>; IETF Tokbind WG <[e-mail]>
Subject: Re: [Unbearable] Possible attack on Token Binding with RSA key exchange

I agree that at this point it doesn't make sense to limit TB to (EC)DHE.

I'm fine with writing a PR to cover this, but I'm unsure what to put in it. I liked how draft-balfanz-tls-channelid started its security considerations by describing 4 classes of network attackers, and I'd like to do something similar here. However, it's unclear to me which of those classes of network attackers are considered as part of the threat model for Token Binding. Would it be reasonable to say that defending against attackers with possession of a server's private key is outside of the scope of Token Binding?

On Mon, Sep 4, 2017 at 3:59 PM, Andrei Popov <[e-mail]> wrote:
> It seems clear that binding tokens to broken TLS channels does one no 
> good, but I don’t mind adding some statements to this effect in the 
> security considerations.
>
> The statements would need to be general enough to cover TLS client and 
> server compromise, PFS and non—PFS.
>
>
>
> Nick, please feel free to author a PR. If we get to IESG review 
> someday, there will be more edits, and we can incorporate this as well.
>
>
>
> I don’t think it makes sense to limit TB to (EC)DHE, especially if 
> we’re going to allow TB in combination with TLS 1.3 0-RTT mode.
>
>
>
> Cheers,
>
>
>
> Andrei
>
>
>
> From: Unbearable [mailto:[e-mail]] On Behalf Of 
> Subodh Iyengar
> Sent: Friday, September 1, 2017 3:28 PM
> To: Nick Harper <[e-mail]>; IETF Tokbind WG 
> <[e-mail]>
>
>
> Subject: Re: [Unbearable] Possible attack on Token Binding with RSA 
> key exchange
>
>
>
> Having said that, even if we only support forward secure ciphers, I do 
> not think we can reasonably claim that we are secure against the 
> threat of stolen private keys.
>
> I like the the carrot opportunity this brings to encourage adoption of 
> forward secure cipher suites.
>
>
>
> Subodh
>
> ________________________________
>
> From: Subodh Iyengar
> Sent: Friday, September 1, 2017 3:24:25 PM
> To: Nick Harper; IETF Tokbind WG
> Subject: Re: [Unbearable] Possible attack on Token Binding with RSA 
> key exchange
>
>
>
> +1 to requiring that token binding only be used with forward secure 
> +cipher
> suites.
>
> I believe the extended master secret requirement of token binding 
> prevents the attack on forward secure suites, i.e. an attacker cannot 
> change the DH share advertised by the real server and thus needs to 
> get the private key for it.
>
> Subodh
>
> ________________________________
>
> From: Unbearable <[e-mail]> on behalf of Nick 
> Harper <[e-mail]>
> Sent: Friday, September 1, 2017 3:03:37 PM
> To: IETF Tokbind WG
> Subject: [Unbearable] Possible attack on Token Binding with RSA key 
> exchange
>
>
>
> I came across an attack on Token Binding today, which I think is worth 
> addressing in TBPROTO in some fashion (likely another paragraph in 
> Security Considerations). This attack involves an adversary with the 
> private key of a server it wishes to impersonate, and was mentioned in 
> draft-balfanz-tls-channelid in the last paragraph of its Security 
> Considerations.
>
> In brief, if an attacker has possession of a server's private key, it 
> can hijack a TLS connection between client and server if the 
> connection uses RSA key exchange instead of (EC)DHE key exchange, 
> which allows the attacker to exercise the bound token without 
> possession of the Token Binding private key.
>
> I realize that we're past WGLC at this point, but I think this should 
> be addressed. On one end, we could require forward-secret key exchange 
> modes with Token Binding. We could also describe this specific attack 
> in the Security Considerations, or we could expand the Security 
> Considerations to describe what attacks are and aren't in the Token 
> Binding threat model, to say that attacks where the adversary has the 
> server's private key are out of scope.
>
> _______________________________________________
> Unbearable mailing list
> [e-mail]
> https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Furlde
> fense.proofpoint.com%2Fv2%2Furl%3Fu%3Dhttps-3A__www.ietf.org_mailman_l
> istinfo_unbearable%26d%3DDwICAg%26c%3D5VD0RTtNlTh3ycd41b3MUw%26r%3Dh3J
> u9EBS7mHtwg-wAyN7fQ%26m%3DaK_8DMmoWWiiQJmFt3JvhVLT6eh7Qm3nf3WuMhMC7vM%
> 26s%3DB0sKQEO8i8hDpLizsh7X8ins9EGGwOwcckYqZku7zEE%26e&data=02%7C01%7CA
> ndrei.Popov%40microsoft.com%7Cc4377a40e8734b28e71808d4f4bd73ea%7C72f98
> 8bf86f141af91ab2d7cd011db47%7C1%7C0%7C636402541906078858&sdata=l4HHrX6
> 7T4Dn1FVfBHlw1xVUv%2B3BSJ867Pf5ezGF3%2B0%3D&reserved=0=