- Test - lists.nanog.org

ATM thruput [was Re: US Domain -- County Delegations]
by kwe＠6SigmaNets.com 02 Aug '95

02 Aug '95

At 6:40 PM 8/1/95, Sean Doran wrote: > >Say, does anyone know if ATM NAP gear can handle > 20Mbps yet? > > Sean. Sean; I have recent preliminary test results of the cisco AIP across the Newbridge 36150 (which is the PacBell NAP configuration) that show nearly full bandwidth TCP throughput. The Kentrox ADSUs are much improved, but results are preliminary. I will present more about this next Tuesday at the PacBell NAP Users meeting in San Ramon, CA, at the next NANOG and eventually on the PacBell NAP Home Page. --Kent

2 1

Re: US Domain -- County Delegations
by Paul A Vixie 02 Aug '95

02 Aug '95

Someone asked me privately... > One thing to keep in mind is that when thinking about appropriate >domain names for a big business, say AT&T, it makes very little sense to >force someone to connect to somemachine.att.nj.us, especially when that >machine could be anywhere in the US or even outside the country. For a >company that large, it would be better to hide geographic info behind a >.com, than to force the geographic location of the company's US >headquarters to influence the domain names of the rest of the company. I'm >sure the folks at IBM would feel the same way. So would all the companies >that are incorporated in Delaware for tax reasons, but have their HQs in >other states. > > For a smaller company, or one with less offices, it might make >sense for their domain name to be geographical, for example cisco.sf.ca.us >or nytimes.ny.us. > > When it comes down to it, I'm agreeing with you, but for a slightly >different reason. It seems a little impractical for the internet to be the >great equalizer between Joe's Garage and Ford Motor Company, especially >when there may be twenty different Joe's Garages but there is only one FMC. i do not believe any more that a company's (or organization's or person's) domain name ought to map differently based on their geographic domain. if we allow that binding, every midsized company who wants to look like a big one will find some way to bend the rules. my current thinking is to match costs and values. folks seem to value having their name directly under .COM, and lo and behold it costs the 'net more to have those names there than if they were all delegated to smaller registries. ergo, the folks who want to live in .COM should probably pay more -- a lot more -- than the folks who don't. extending this model upward, it should be poss- ible to register directly in the root, for US$1M a year or so. the idea here is, let the company categorize themselves. no rules to bend. but: if they want to live in a place that's expensive for the 'net to maintain, they should pay for the overhead they are generating. as for all those deleware corporations, i'd expect each company to choose the local registry that makes the most sense -- to them, not to me or to the NIC or the IANA. if they are incorporated in deleware but have their main offices in virginia and want to register their domain under alaska, that's noone's business but their own. the alaskan registry would be within its rights to charge more for out of state registrations.

4 6

Re: US Domain -- County Delegations
by kwe＠6SigmaNets.com 02 Aug '95

02 Aug '95

At 4:16 PM 8/1/95, Walter O. Haas wrote: >My thinking has evolved away from geographically based names... >the Internet is >erasing geography, and I think that we would be living in the past if >we relied too much on geography in naming. It might make sense for >"joes-garage" to be a geographically-qualified domain since the market >for automobile repair is likely to be within a few miles of the business, >but my little one-man consulting corporation is located in Salt Lake City >with it's biggest customer in St. Louis. For a business with essentially >no geographic barriers, such as mine, it doesn't make much sense to >include geography in the name regardless of where the corporation is >registered. Walt; This is true for non-geographic markets, but what happens when the Internet reaches a size where geographic markets develop? In other words, when it makes sense to buy pizza on the Net, won't it make sense to revive geographic naming, to serve geographic markets? --Kent

3 2

Re: root name servers
by Michael Dillon 02 Aug '95

02 Aug '95

On Tue, 1 Aug 1995, Kent W. England wrote: > At 8:43 AM 7/29/95, Michael Dillon wrote: > > > >What about hijacking the root name servers? Here's the scenario..... > ... > >Could people be plotting this even as we speak? Is it a good idea? Would > >it solve the namespace problems we currently have? > > > >Michael Dillon > > I want bork.bork.bork and die.die.die top level domains, please or I will > hijack your hijack. Thank you. The whole point of hijacking the root domain servers is that it turns DNS from a centralised system into a more distributed system where cooperation is essential. I could be wrong because I don't fully understand BIND internals here, but if anyone wanted to copy the root domain and add new toplevel domains like .bork or .die on their own networks, they could just do it. Any references to .bork or .die would be correctly resolved and any other references would be delegated to the same .com and .org and .net servers that are currently in use. If other networks felt there was value in accessing the .bork and .die domains they would do similarily referencing your nameserver to resolve such references. Is there any technical reason this wouldn't work? Michael Dillon Voice: +1-604-546-8022 Memra Software Inc. Fax: +1-604-542-4130 http://www.memra.com E-mail: michael(a)memra.com

2 2

Re: root name servers
by kwe＠6SigmaNets.com 01 Aug '95

01 Aug '95

At 8:43 AM 7/29/95, Michael Dillon wrote: > >What about hijacking the root name servers? Here's the scenario..... ... >Could people be plotting this even as we speak? Is it a good idea? Would >it solve the namespace problems we currently have? > >Michael Dillon I want bork.bork.bork and die.die.die top level domains, please or I will hijack your hijack. Thank you. --Kent :-)

1 0

Re: US Domain -- County Delegations
by rrv＠uiuc.edu 01 Aug '95

01 Aug '95

At 2:39 AM 8/1/95, Paul Traina wrote: >Just to be obnoxious, let me state the ugly obvious programatic mapping... >use the first 2 or 3 letters in the 2nd layer domain name. > >cisco.com = cisco.ci.com >nytimes.com = nytimes.ny.com >real-routers.com = real-routers.re.com > >Since this is a purely implied and programatic mapping, the particularly >-cute- thing we can do is automagicly map this crud so no user ever has >to see the abomonation "cisco.ci.com". Better yet, hash the 2nd layer domain name.

6 8

fyi - routing stability reports
by Curtis Villamizar 01 Aug '95

01 Aug '95

FYI- Did anyone notice they weren't getting routing stability reports from ANS? Don't answer that, I get enough mail as it is. :-) Due to a crashed disk, we lost data between July 8 when the last backup on this non-essential filesystem occurred and August 1 when we replaced the drive. You may have noticed lack of reports since July 17 (prior to that reports were generated, but we hadn't archived them to tape, so we've lost them. The main reson I'm asking is that I may have added report generation for one or two AS that don't border us and if this was between July 8 and July 17 the addition has been lost so you have to make the request again. Curtis ps - the routing stability reports are not an officially supported service of ANS. They exist to help identify operational problems. Other providers find them useful, so I'm trying to keep them going.

1 0

Re: US Domain -- County Delegations
by Paul A Vixie 01 Aug '95

01 Aug '95

a private conversation has borne some useful fruit: >> I gathered from some mail that Paul sent out recently >> that there was no problem with overloading of the root >> servers. Something about the fact that his BSD/OS Pentium >> was handling 100lookups/second without breathing hard... > The question is is the growth exponential? the 100/sec number has been pretty stable for the last four months. it was 80/sec last year. it bursts up to 150/sec a few times an hour. no big deal. i could do 600/sec before i would have to worry about the cpu speed. i would fear for the interconnects if all seven root servers got 600/sec. there are two kinds of root server queries. one is a name server starting up and trying to "bootstrap" its cache using its root.cache or named.ca or what- ever file. the other is when a recursive server doesn't have an answer and has to delegate. therefore the root server load scales with the number of recursive name servers, not with the number of clients or the number of domains registered. if it were otherwise, the world would have melted long before now.

1 0

BMWG/IPPM Meeting Adjacent to NANOG
by almes＠advanced.org 01 Aug '95

01 Aug '95

As you develop your travel plans for the Pittsburgh NANOG meeting (which runs from morning of Monday, 11-Sep, through noon of Tuesday, 12-Sep), be aware of the following. At the Stockholm IETF meeting, there was a decision to have an interim meeting of the BMWG regarding our new effort in the area of IP Provider Metrics, and that the best time to have this meeting would be adjacent to NANOG. So, a more formal message defining that meeting and its agenda will be sent to the appropriate lists within a week. For the present, I only want to encourage anyone who will be attending NANOG and is interested in IPPM agenda to reserve the afternoon of Tuesday, 12-Sep, between 1:30p and 5:00p, at PSC, for that meeting. -- Guy

1 0

Re: root name servers
by Robert E. Seastrom 01 Aug '95

01 Aug '95

> Erik Sherk and I have arranged for internic to be one hop off of > suranet's router on mae-east. Additionally, expect to see some changes > in the root servers configuration on namedroppers as soon as I can dig > out of my large amount of mail that I recently have received... terp.umd.edu will likely be in a similar situation real soon now, but i'm not the one who's coordinating that. ---Rob

1 0