GPRS. domain Root server

geert-jan not at valid.com
Mon Jul 26 12:08:29 UTC 2004


<phn at icke-reklam.ipsec.nu> wrote in message
news:cdiok7$2mdt$1 at sf1.isc.org...
> Brook Harty <harty at ironwolve.com> wrote:
> > Anyone know if there is a GPRS root domain server? British telecom had
> > the primary and secondary servers up, but have decided to stop
> > providing this service. They don't want to be a GRX provider anymore.
>
> No. It has been suggested, but rejected. It's unclear what benefits
> it should give the world,

It was conceived for 'GPRS network internal' resolution. The mobile message,
originating from a mobile terminal, sent to a terminal that is not hooked up
to the local telco's GPRS network needs to be relayed to the outer boundary
of that GPRS network (this functionality was called APN routing).  This
outer boundary is a node called GGSN. So the 'DNS for GPRS' was setup to
find the proper GGSN. The GGSN is a router that could either relay the
message to the proper telco (through the GRX name resolution) or to the
internet.

> > What is the possibility about getting GPRS. into the normal Internet
> > root servers? (Or is this Crazy talk?)
> None. Better educate admins in the telecom industry about sane and
scalable
> practices.

There are very good admins in the telecom industry. But organisations like
3GPP should come up with better, more scalable standards. And when using
established standards, they should not violate against them, like Jim Reid
eloquently explained. I agree with other people in this newsgroup that the
whole setup is in clear violation of RFC2826 - "IAB Technical Comment on the
Unique DNS Root".
For the record, the IETF has requested the 3GPP to NOT further extend the
use of the ".gprs" domain.

> For instance the phrase "GPRS" comes from a technical implementation of
> one way to implement mobile phone systems. And as such short-lived and
> transient. What is the problem they try to solve by using "gprs" ?

See explanation above.

--
Geert-Jan




More information about the bind-users mailing list