RE: [Dime] Issue39: What to verify in TLS certificates

["Asveren, Tolga" <tasveren at sonusnet.com>]

Hi Victor,

I think we need to agree on some form of post check for TLS certificates
for interoperability purposes. AFAICS, there is nothing we need to
define new from technology perspective; we just need to decide the
fields to use in the certificate and the meaning of the corresponding
values. Overall, this probably will be a few paragraphs.

I don't know what is the best place to define this, just thinking a
separate document only for this issue could be an overkill (I don't know
whether there are plans for a security type of document).

    Thanks,
    Tolga

-----Original Message-----
From: Victor Fajardo [mailto:vfajardo at tari.toshiba.com] 
Sent: Monday, April 16, 2007 11:13 AM
To: Asveren, Tolga
Cc: dime at ietf.org
Subject: Re: [Dime] Issue39: What to verify in TLS certificates

Hi Tolga,

Comments inline:

> If we use subjectAltName for a postcheck for TLS certificates in the
following way:
> a) dNSName contains the name of the domain, the host claims to be in
(OriginRealm in CER/CEA)
> b) otherName contains the Id of the application, which the host claims
to host for the realm specified in a)
>
> Do we need to verify that the TLS certificate is valid for the domain
of the server which was returned by the DNS query as well? It doesn't
seem necessary to me, afterall what we care is that the host is valid to
provide a given service for a given domain.
>   

My personal preference is not to even deal with this issue in the base 
protocol simply because certificate ownership/usage validation is may 
not be limited to diameter only. The "issue" may also be relevant to 
other protocols so its probably better to "specify" the solution (if 
there isn't already one) in more relevant security related documents 
which the base protocol can simply refer to.

> As an example if we are interested in getting serviceA for domainA and
if at the end of DNS queries we end up with
host5.thirdpartydiamnetwork.com, are we really interested that we have a
valid certificate for the domain thirdpartydiamnetwork.com?
>
> Currently, in RFC3588 we have the following:
>
> If the server is using a site certificate, the domain name in the
query and the domain name in the replacement field MUST both be valid
based on the site certificate handed out by the server in the TLS or IKE
exchange. 
>   

The example you mentioned above re-enforces my idea so I agree that it 
maybe better to remove these text from the base spec. It also helps 
simplify the base protocol similar to the removal of the IPsec usage.


regards,
victor

> Does that text refer to the type of usage I think as unnecessary?
>
>    Thanks,
>    Tolga 
>
> _______________________________________________
> DiME mailing list
> DiME at ietf.org
> https://www1.ietf.org/mailman/listinfo/dime
>
>
>
>   


_______________________________________________
DiME mailing list
DiME at ietf.org
https://www1.ietf.org/mailman/listinfo/dime