ietf
[Top] [All Lists]

Re: Last Call: draft-daboo-srv-caldav (Use of SRV records for locating CalDAV and CardDAV services) to Proposed Standard

2010-06-23 14:02:50
Basically, yeah, as long as you have DNSSEC. In fact, if you've got DNSSEC, you don't really even need the application-specific bit at the end. The goal of the XMPP DNA work (and other analogous things) is to work around not having DNSSEC in the mean time. In that solution, the parallel path is:

- Announce the fact that example.com is hosted at clendarserverfoobar.com in DNS

- Connect to calendarserverfoobar.com over TLS and check that the name in the cert is indeed calendarfoobar.com

- Obtain an attribute cert for calendarfoobar.com signed by example.com

- Valid signature and certificate for example.com implies that delegation is OK

The third of these steps of course require some application-specific magic.

--Richard


On Jun 23, 2010, at 2:52 PM, Patrik Fältström wrote:

On 23 jun 2010, at 16.33, Richard L. Barnes wrote:

In principle, example.com is the proper domain to authenticate, but in practice, that causes a lot of problems. Consider the case where the target of the redirection is a separate entity from the origin; this could arise, for example, in a situation whereexample.com has outsourced its calendaring services to calendardserverfoobar.com.

So, the "connect the dots" is to:

- Announce the fact example.com is hosted at calendarserverfoobar.com (with some URL) in DNS

- Secure that announcement in DNS with DNSSEC

- Verify the SSL (for example) cert for the connection to calendarserverfoobar.com matches

- Do application layer authentication etc over the then encrypted connection

Sounds ok?

  Patrik


_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf

<Prev in Thread] Current Thread [Next in Thread>