To:
marka@isc.org
Cc:
mohta@necom830.hpcl.titech.ac.jp, Mark_Andrews@isc.org, randy@psg.com, dnsop@cafax.se
From:
Masataka Ohta <mohta@necom830.hpcl.titech.ac.jp>
Date:
Fri, 9 Jul 99 15:08:37 JST
In-Reply-To:
<199907090610.QAA12682@bsdi.dv.isc.org>; from "marka@isc.org" at Jul 9, 99 4:10 pm
Sender:
owner-dnsop@cafax.se
Subject:
Re: Topological Motivation for draft-ohta-root-servers-01.txt?
Mark; > > > Can the existing root nameservers cope with however many million > > > servers there are in the world all deciding within the refresh > > > interval to transfer a copy of the root zone? > > > > Of course. > > Then lets do it and put this draft in the bit bucket!!!! It is so obvious that it does not worth spending lines. > It's easy enough to make nameservers choose a root server > at random to transfer from:-) Urrr, it is certainly true that the average diameter of random graphs is still logarithmic to the number of nodes. However, the degree of each node is not bounded. Worse, there is a small possibility of graph separation. So, it's better to use more systematic approach based on a tree with additional redundant edges. > > > Every country needs good root name service. > > > > Every country needs good root name service at their definition > > of the goodness. > > Good to hear that you agree with me that it is service > level that is driving this and not politics. What, do you think, defines the service level? Hint: It is not technology. Masataka Ohta