uDNS Root Name Servers Taking Shape - on a couple ISDN lines
Ben Black
black at zen.cypher.net
Fri May 30 00:24:41 UTC 1997
kook fight in the parking lot at 3 o'clock after school!
On Fri, 30 May 1997, Ron Kimball wrote:
> On Thu, 29 May 1997 18:03:13 -0500, Karl wrote:
>
> >> ;; ADDITIONAL RECORDS:
> >> root.starfire.douglas.ma.us. 86400 A 208.195.108.8
> >Multi-homed condition unknown and suspect due to truncated BGP path.
>
> Yup, not multihomed until the new router comes in. :-(
>
> >Approximate bandwidth from the core to this point on the network
> >from us at this point in time: 34.56kbps, or a good modem line :-)
>
> Gee, it's a T1 from here, must be a problem on your end. <grin>
>
> >THIS NAMESERVER IS RUNNING WITH RECURSION ENABLED
>
> Yup, until next week when we get the new box up.
>
> > AND IS NOT A TRUE ROOT.
>
> Says you, the grand high holy keeper of the ONE TRUE ROOTS. Ha!
>
> >> hp.manhattan.com. 172800 A 199.103.194.137
> >Aggregated by (and complete path from) Open Advisors. Appears to be
> >multi-homed.
>
> Yup.
>
> >Approximate bandwidth to this point on the network: 65.28kbps, or a
> > single-channel ISDN equivalent.
>
> You really should check your lines Karl, a multihomed server on a
> single channel ISDN, I don't think so...
>
> >**** NOTICE:
> >THIS NAMESERVER IS RUNNING WITH RECURSION
>
> Hmm... the name.boot file has it set off. I'll check it out.
>
> AND IS NOT A TRUE ROOT.
>
> <yawn>
>
> >> DONTSERF.MAKEWAVES.NET. 172800 A 204.94.43.1
> >Alternic under a different name, operated by Diane Boling, and running
> >with both nameservers on the same subnet. Linked to Seanet, which appears
> >to be multihomed.
>
> Yup.
>
> >Approximate bandwidth to this point on the network: 629kbps (my god, they
> >have one root with a fractional T1 worth of bandwidth available!)
>
> Well, I guess your lines came back up! <grin>
>
> >**** NOTICE:
> >THIS NAMESERVER IS ALSO RUNNING WITH RECURSION ENABLED
>
> could be.
>
> AND IS NOT A TRUE ROOT.
>
> <yawn>
>
> >I rest my case. Only one of these has anything approaching reasonable
> >connectivity, all appear to be off single-point failure circuits (except
> >possibly manhattan.com), and all are running in non-RFC2010 mode.
>
> Yah, we really need RFC2010 servers to run 1/2% of the internet - NOT!
>
> Seriously, our schedule calls for 5 dedicated, non-recursive servers
> up by next week this time, with T1 of better connectivity. We plan
> full RFC2010 by the time we reach 5% visibility. Feel free to market
> your system's RFC2010 compliance as an absolute must for servers that
> handle a fraction of a percent of the internet's DNS requests, I'd be
> surprised if any of the "internet aware" people on these lists you are
> posting to care...
>
> Take care,
> Ron Kimball for the uDNS council
>
More information about the NANOG
mailing list