(Non)-Uniqueness of Net Names
Any strong opinions? I'll answer the factual questions. Daniel ------- Forwarded Message Date: Mon, 05 Jul 1993 23:14:10 -0400 From: Mark Knopper <mak@merit.edu> Resent-From: RIPE NCC Staff <ncc@ripe.net> To: ncc@ripe.net, merit-ie@merit.edu, conf@merit.edu, dsj@merit.edu, markk@internic.net, scottw@internic.net Resent-To: RIPE NCC Staff <staff@ripe.net> Subject: (Non)-Uniqueness of Net Names Hi. I'm including the RIPE and InterNIC folks on this discussion. Here are some questions for your consideration: 1) Should we remove the uniqueness requirement for net names in the Merit and InterNIC/whois databases? 2) My understanding is that RIPE's database does not require existence or uniqueness of net names. Is that correct? 3) Do net names serve any useful purpose? Should we get rid of them altogether? If we reach consensus here, I suggest that we issue an Internet draft containing a proposal. If no major concerns are raised we can make the decision final and issue an RFC. What do you think about this suggestion? What other people or organizations should be consulted in this process? Mark .... included messages omitted ------- End of Forwarded Message
I don't find that net names are of very much use. If there is to be no requirement for uniqueness, they become even less useful. If they are to be preserved and to be unique, then some quite meaningless names will emerge. Also, to guarantee uniqueness would be quite a chore - but that's an administrative detail ;-) Cheers. Mike
In message <9307060818.AA18021@ncc.ripe.net>, Daniel Karrenberg writes:
Any strong opinions? I'll answer the factual questions.
Daniel
------- Forwarded Message
Date: Mon, 05 Jul 1993 23:14:10 -0400 From: Mark Knopper <mak@merit.edu> Resent-From: RIPE NCC Staff <ncc@ripe.net> To: ncc@ripe.net, merit-ie@merit.edu, conf@merit.edu, dsj@merit.edu, markk@internic.net, scottw@internic.net Resent-To: RIPE NCC Staff <staff@ripe.net> Subject: (Non)-Uniqueness of Net Names
Hi. I'm including the RIPE and InterNIC folks on this discussion. Here are some questions for your consideration:
1) Should we remove the uniqueness requirement for net names in the Merit and InterNIC/whois databases?
2) My understanding is that RIPE's database does not require existence or uniqueness of net names. Is that correct?
3) Do net names serve any useful purpose? Should we get rid of them altogether?
We use the net names to produce a much more readable IP accounting/ statistics from the routers. So in my opinion they serve a useful purpose and should not be thrown away. -- Arne
1) Should we remove the uniqueness requirement for net names in the Merit and InterNIC/whois databases?
Yes yes yes yes yes yes yes please ! Merit and I have had a number of headaches trying to get things straight in their database when I've been registering blocks of C networks for NSFnet access, when the block has essentially just had one name for the entire block. Even if two nets have the same name and belong to different organisations, the other information you pull back from the database(s) should clear up any confusion.
3) Do net names serve any useful purpose? Should we get rid of them altogether?
Certainly, particularly if a block of network numbers can just have a single name. I'd be loathe to have no option but to refer to a network only by its number. Dunc
Daniel Karrenberg writes:
1) Should we remove the uniqueness requirement for net names in the Merit and InterNIC/whois databases?
2) My understanding is that RIPE's database does not require existence or uniqueness of net names. Is that correct?
3) Do net names serve any useful purpose? Should we get rid of them altogether?
The most useful reason for their being there IMHO is to be able to build an /etc/networks file from them. This means, of course, they need to be unique. Regards, Cliff. -- Cliff Stanford Demon Systems Limited Demon Internet Services 42 Hendon Lane *********************** London N3 1TT England THE INTERNET FOR A TENNER-A-MONTH 081 349 0063
Cliff Stanford <cliff@demon.co.uk> writes:
The most useful reason for their being there IMHO is to be able to build an /etc/networks file from them. This means, of course, they need to be unique.
I thought we were beyond flat file directories *on a global scale*. Daniel
The most useful reason for their being there IMHO is to be able to build an /etc/networks file from them. This means, of course, they need to be unique.
I thought we were beyond flat file directories *on a global scale*.
Not really, we use the /etc/networks file for our IP-accounting as a last resort if part of the DNS system is out of order. Martijn.
participants (6)
-
Arne Asplem -
Cliff Stanford -
Daniel Karrenberg -
Duncan Rogerson -
martijn＠nluug.nl -
Mike Norris