Re: [DNSOP] New Version Notification for draft-pusateri-dnsop-update-timeout-00.txt

[Ted Lemon <[e-mail]>]

Sorry, I realized that I accidentally hit "reply" instead of "reply all."

The issue that I raised with Tom is that for the DNSSD SRP use case, the only names that receive updates from multiple services are service names (IOW, not service instance names).   In the case of SRP, PTR RRs in service names always point to service instance names, which are per-service, and hence can be counted on to expire on a single schedule.   So for the use case of SRP, the easiest way to handle deleting service name PTR RRs is to take advantage of the semantics of DNSSD: when a service instance SRV record expires, also delete the PTR RR on the service name that points to it.

The reason I bring this up is that it's the most complicated use case I know of.   Is there some other use case where we expect more than one DNS Update client to be updating RRs of the same type on the same name?   How would this even work without SRP semantics?   If this is not expected, then any complexity in the timeout RR that's present only to support that use case is unnecessary.

This is what has been motivating my questions about use cases.   I'm sorry I didn't make that clear earlier.

> On Aug 27, 2018, at 3:34 PM, Tom Pusateri <[e-mail]> wrote:
> 
> Not true. You have PTR records with the same service name from multiple clients each with different RDATA and unique timeouts as I demonstrated yesterday in the example.
> 
> Tom
> 
> 
>> On Aug 27, 2018, at 3:27 PM, Ted Lemon <[e-mail] <mailto:[e-mail]>> wrote:
>> 
>> For service instance names, there is only one service. So there shouldn’t be a problem. 
>> 
>> On Mon, Aug 27, 2018 at 2:28 PM Tom Pusateri <[e-mail] <mailto:[e-mail]>> wrote:
>> Because even if you add TYPE, you have multiple PTR records (instance names) with the same owner name, class, and type that can timeout differently.
>> 
>> Tom
>> 
>>> On Aug 26, 2018, at 11:20 PM, Ted Lemon <[e-mail] <mailto:[e-mail]>> wrote:
>>> 
>>> If we do that, why do we need a hash at all?
>>> 
>>> On Sun, Aug 26, 2018 at 10:51 PM Mark Andrews <[e-mail] <mailto:[e-mail]>> wrote:
>>> I would add a covered type field to TIMEOUT (c.f. RRSIG).  I also wouldn’t have more than
>>> a single timeout per record.  I’m tempted to say a single hash as well.  If there is multiple
>>> timeouts per record then the blocks need to be sorted in timeout order.
>>> 
>>> Covered is there to reduce the number of RR’s that need to be hashed to remove a record.
>>> It will also reduce the size of IXFR’s as you don’t need to re-construct a new TIMEOUT
>>> record that covers every timeout at a name on each change.
>>> 
>>> For all records at a name is often more expensive that for all records of type covered.
>>> Name servers are optimised for looking up <name,type,class> tuples rather than <name,class>
>>> tuples.
>>> 
>>> Sorting of timeout blocks is so that you can look at the first timeout when working out
>>> which TIMEOUT needs to be processed first in a zone.
>>> 
>>> -- 
>>> Mark Andrews, ISC
>>> 1 Seymour St., Dundas Valley, NSW 2117, Australia <https://maps.google.com/?q=1+Seymour+St.,+Dundas+Valley,+NSW+2117,+Australia&entry=gmail&source=g>
>>> PHONE: [phone]              INTERNET: [e-mail] <mailto:[e-mail]>
>>> 
>> 
>> _______________________________________________
>> DNSOP mailing list
>> [e-mail] <mailto:[e-mail]>
>> https://www.ietf.org/mailman/listinfo/dnsop
>