Re: in-addr.arpa zone data in RIPE database
- Date: Wed, 03 Mar 93 12:18:35 +0100
> jalm@localhost (Jose Legatheaux Martins) writes:
> Conclusion: we should invest in tools to detect such errors and
> notify the administrators concerned. There are some of these
> tools available in the archive of RIPE. Anyway, I, as an administrator,
> would prefer to have a single point of focus for stating which
> are the servers running authoritative for my domain or for
> the reverse mapping of my network and that is the DNS, nothing more.
Fully agree and the proposal is to use the RIPE DB for this in Eurpe.
> 2) We have the RIPE database: this is a very good tool to deal
> with some administrative data. Well, lets use it for this purpose.
> Anyway, we should avoid to have to deal manually with DUPLICATE
> data. So, we should avoid to have information about servers
> and addresses of those servers in the RIPE dbm.
Of course there is a chicken and egg problem here. In order to use the
DNS for anything, the appropriate NS RRs have to be put in somewhere.
One has to keep a record of which NS RRs go where. Currently
in-addr.arpa is centralised and this means that we need a centralised
registry.
> My suggestion is: use the RIPE dbm to deal * only * with those
> informations that cannot be effectively registered in the DNS.
Wholeheartedly agree. Remember I proposed to get rid of the domain
object in the RIPE DB? Well I was told that there is some information
in those objects that connot be represented in the DNS. So we kept it.
> If we decide to maintain in the RIPE dbm information about
> DNS servers, those fields should be automatically updated from
> the DNS on a (for example) monthly basis.
See "chicken and egg" above.
> I do not believe on centralized repositories and management unless
> I have no other way to avoid it.
In general you are right.
> I do not believe in consistency maintained mannually by hundreds
> (thousands ?) of administrators.
That's why you are right only "in general".
Daniel
|