> Suppose "nslookup" does domain name to adress translation,
> "net" does IP adress to IP net translation and finally
> "connect" gives the connect information about IP networks
What I was thinking of was the "*n[io]" - tag already used.
(e.g: *ni: 1=701 2=1800 3=1238 )
> (This information already is in the RIPE DB), then
> connect ( net ( nslookup ( domain name ) ) )
> does what you want.
>No, it doesn't. The connect field gives the connectivity
>a given network has, partly in "real" connectivity (e.g.
>"EU"), partly in "pseudo" connectivity ("RIPE"), but it
>doesn't state which the network service providers are nor
>nor the order in which these are service providers. And
>the latter is extremely important for determining policy
>based routing. The least that should be done to treat the
>connect field as routing source is to list the mnemonics
>that stand for the real service providers in order of
>routing preference (for the associated network of course).
>However, I don't like this "solution" at all, given the
>mix of "real" and "pseudo" connectivity in the connect
>field. Besides, a new field could be based an AS numbers
>in order of preference, much like the NSFnet policy routing
>database does now, making life infinitely easier.
>
>Furthermore I would consider building potentially large
>access or other lists for a router in the suggested way,
>through massive DNS queries, a crime...
As far as I remember there is a "*di" - tag for domain name entries
listing IP adresses used in this domain. Hence no usual "nslookup"!
>
> Don't overload the DB with information (Current size: 7315558 Byte.
> We started with some hundred KB).
>It will keep on growing anyway, especially when the RIPE,
>NIC and Merit databases get synchronized. Database size
>should not be used as an argument to keep out *useful*
Don't misunderstand me!! I don't want to keep out *useful* information.
I just want to minimize and streamline the DB. Keep the DB as small as
possible and join the information on your local machine (NO DNS lookups!).
>and authoritative information.
>
>
> Piet
Arnold
