control channel protocol?

[Mark Andrews]

In message <4CBC3A5F.6020501 at knipp.de>, Klaus Malorny writes:
> On 18/10/10 01:16, Doug Barton wrote:
> > On 10/16/2010 4:58 AM, Klaus Malorny wrote:
> >> [...]
> >
> > It's virtually certain that you would get better results with less effort b
> y
> > switching to using dynamic updates.
> >
> >
> > hth,
> >
> > Doug
> >
> 
> Hi Doug,
> 
> well, I considered this, but as far as I understand, there are some non-trivi
> al 
> problems with RFC 2136:
> 
>   - can't create or drop zones -> still need to use the classic approach
> 
>   - updates limited to a single message (64k limit via TCP), requiring
>     the updates to be split if need be -> problems regarding atomicity,
>     query integrity during update, serial numbers

A 64K update is a really, really large update.  The unsigned root
zone fits into 64K.  I know automated systems can generate really
large updates but in reality almost all of those can be split up
safely especially it the automated system is generating the update
incrementally.

> also, this is only considered as an interim solution until Bind10 becomes 
> suitable for our needs.
> 
> Regards,
> 
> Klaus
> _______________________________________________
> bind-users mailing list
> bind-users at lists.isc.org
> https://lists.isc.org/mailman/listinfo/bind-users
-- 
Mark Andrews, ISC
1 Seymour St., Dundas Valley, NSW 2117, Australia
PHONE: +61 2 9871 4742                 INTERNET: marka at isc.org