What about something like this:

When using BIND, or other software that can act as both a recursive and authoritative server in the same instance, there is a tradeoff between using a separate view (or separate instance) for slaving the root zone, versus slaving the zone into the same view (or instance) where the recursion is occurring.

Using a separate view/instance has the advantage that when the recursor is also performing DNSSEC validation that the DS records in the slaved zone will also be validated. In BIND this validation does not occur when the zone is slaved into the same view/instance where the recursion/validation is occurring.

Slaving the zone into the same view/instance as the recursion has the advantage that when changes happen to the data in the zone the recursive view/instance will be updated as soon as it receives its copy of the zone. When using a separate view for slaving the zone the recursive instance will cache all of the queries it looks up. Currently the TTL for DS and delegation NS records is 2 days.


On 11/20/14 10:52 AM, Bob Harold wrote:
Thanks Paul,
    I use BIND, but am not an expert.  Based on the discussion I will
suggest some words and the experts can correct me:

    Note:  By using a separate view, the "recursive" view will do DNSSEC
    validation on the responses it receives from the "root" view, which
    is necessary for security.  It will cache the answers, including the
    validation.

    Alternatively, if the root zone was loaded directly in the
    "recursive" view, then DNSSEC validation would not be done, as BIND
    would trust the zone.  Then you would want to do separate validation
    on the zone during zone transfers.  This might result in less
    caching and less time spent validating, but requires a more complex
    configuration.




--
Bob Harold
hostmaster, UMnet, ITcom
Information and Technology Services (ITS)
rharo...@umich.edu <mailto:rharo...@umich.edu>
734-647-6524 desk

On Thu, Nov 20, 2014 at 1:25 PM, Paul Hoffman <paul.hoff...@vpnc.org
<mailto:paul.hoff...@vpnc.org>> wrote:

    On Nov 20, 2014, at 10:20 AM, Bob Harold <rharo...@umich.edu
    <mailto:rharo...@umich.edu>> wrote:
    > I can see where "validate on zone transfer" would be a feature request.  And 
"validate everything" similarly.
    >
    > For the draft, could a small paragraph be added explaining the difference 
between using a separate view for the root zone and just loading it in the same 
view, so that people like me realize the tradeoffs before we decide to implement 
the draft with what we might think is a minor simplification, not realizing the 
impact?

    Yes, we can add this to the BIND example in the appendices. Given
    that I kinda suck at BIND, proposed wording would cause less grief
    in the next draft...

    --Paul Hoffman



_______________________________________________
DNSOP mailing list
DNSOP@ietf.org
https://www.ietf.org/mailman/listinfo/dnsop

Reply via email to