Article View: aus.net.policy
Article #581Re: DNS Registry
From: Geoff Huston
Date: Tue, 28 May 1996 00:00
Date: Tue, 28 May 1996 00:00
53 lines
2850 bytes
2850 bytes
> : > Section 4 - Conclusions > : > ----------------------- > > : I would suggest thatyou need to think through the rationale and needs > : very carefully first before leaping off into corporate structures, > : charging, equipment and other paraphenalia. > > Shrug. The idea was to present a complete proposal and let people > rip it up. Having a conjectural "idea" won't see anything done in > a realistic timeframe. I guess my approach to this problem is now readically different than thatof some 18 months ago. There are two basic responses to issues of administration of public resources: centralism with a single public administration or dispersion of the role. Banking exhibts both aspects - a single "central" reserve bank and a disperse retail banking structure. The dispersed structure exhibts a degree of robustness in its diversity and the central structure is heavily regulated in what it can do so that its effects on the retail banking sector are heavily damped. Now this model of banking took some 400 years of financial crises to evolve to what we have today. Is it a good system? Probably, at least in so far as the system has both the confidence of the clients of the system (if you weren't confident in it you wouldn't use it), and it appears to work most of the time (financial crashes are not a regular daily occurrence!). While I'm not suggesting that we ponder the DNS admin issues for the next 400 years I am suggesting that centralism and single point administration may not be the most appropriate answer, nor is a complete dispersed anarchy veryt functional either! I believe we do have the technology and the intelligence to devise a distributed system where there are a number of paths into the DNS - the dispersion of processing and associated administration paths with the constraint of a relatively consistent and stable underlying common policy environment are, in my view, the optimal objectives in this area. Geoff ---- NO permission is granted for the reproduction of any parts of this posting that have been authored by myself in any way other than to be quoted in electronically posted followups of this article. This posting does not represent itself as being the views of my employer. ----
Technical Details
Message-ID:
<31AA5E67.41C67EA6@telstra.net>Path:
arc1-ca.pugleaf.net!unknown!not-for-mailReferences:
<4o6v1t$f5o@folk.iinet.net.au> <31A8F2EF.41C67EA6@telstra.net> <4odc38$nqk@folk.iinet.net.au>Show Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: f4bcb,1dc3422d080819f9 X-Google-Attributes: gidf4bcb,public From: Geoff Huston <gih@telstra.net> Subject: Re: DNS Registry Date: 1996/05/28 Message-ID: <31AA5E67.41C67EA6@telstra.net> X-Deja-AN: 157083822 references: <4o6v1t$f5o@folk.iinet.net.au> <31A8F2EF.41C67EA6@telstra.net> <4odc38$nqk@folk.iinet.net.au> content-type: text/plain; charset=us-ascii organization: Telstra Internet mime-version: 1.0 newsgroups: aus.net.policy x-mailer: Mozilla 3.0b3 (X11; I; SunOS 4.1.3 sun4m)