From owner-netizens-digest@columbia.edu Thu Apr 12 20:25:29 2001 Return-Path: Received: from maillist1.cc.columbia.edu (maillist1.cc.columbia.edu [128.59.35.140]) by mail1.panix.com (Postfix) with ESMTP id 9433D494B7 for ; Thu, 12 Apr 2001 20:25:29 -0400 (EDT) Received: (from majordom@localhost) by maillist1.cc.columbia.edu (8.9.3/8.9.3) id UAA16966 for netizens-digest-outgoing; Thu, 12 Apr 2001 20:21:20 -0400 (EDT) Date: Thu, 12 Apr 2001 20:21:20 -0400 (EDT) Message-Id: <200104130021.UAA16966@maillist1.cc.columbia.edu> From: owner-netizens-digest@columbia.edu (Netizens-Digest) To: netizens-digest@columbia.edu Subject: Netizens-Digest V1 #381 Reply-To: netizens@columbia.edu Sender: owner-netizens-digest@columbia.edu Errors-To: owner-netizens-digest@columbia.edu Precedence: bulk Status: RO Netizens-Digest Thursday, April 12 2001 Volume 01 : Number 381 Netizens Association Discussion List Digest In this issue: Re: [netz] There is a need for online discussion of new DNS NAS Commmittee Re: [netz] There is a need for online discussion of new DNS NAS Commmittee Re: [netz] There is a need for online discussion of new DNS NAS Commmittee ---------------------------------------------------------------------- Date: Thu, 12 Apr 2001 00:56:50 EDT From: Nmherman@aol.com Subject: Re: [netz] There is a need for online discussion of new DNS NAS Commmittee In a message dated 4/11/2001 6:41:26 PM Central Daylight Time, gds@best.com writes: > Furthermore, the ORSC people that I know > aren't purely motivated by profit; arguably, they do want to earn > money for the work they do. > > --gregbo > This is my first time posting to the list. My name is Max Herman and I work with the Genius 2000 Network. If anyone here is offended by me or my presence, please let me know immediately. My comment is how the naming process pertains to a radical transformation of the earth and human habitation. I believe the next decade will be the first decade of massive demilitarization in history. This will cause many current businesses, and businesses dependent on a war economy, to fail in profusion. A large market crash is also indicated in this scenario, as well as political upheaval. Or to clarify, what is the best DNS strategy overall for preserving information diversity (to use a biological expression)? Max Herman The Genius 2000 Network http://www.geocities.com/genius-2000 ------------------------------ Date: Thu, 12 Apr 2001 13:45:57 -0700 (PDT) From: Greg Skinner Subject: Re: [netz] There is a need for online discussion of new DNS NAS Commmittee hcb wrote: > gds wrote: > >I don't understand the point you're trying to make here. Are you > >trying to argue that what new.net is doing is requiring other parties > >(e.g. ISPs) to do something, but new.net is not funding them to do it? > Exactly. Their profit making activity incurs costs on the part of > others who do not share in those profits. Could you clarify what you mean by "requiring ISPs to do something?" What specifically are ISPs required to do, as a result of what new.net is doing? > >I don't think useful comparisons can be made between DNS service (as > >it is currently practiced) and 911 service. [...] > There is a useful comparison, if you accept the idea of cooperation > rather than anyone being free to do their own thing. I can have > confidence that if I dial 911, I will not be connected to a pizza > delivery service. Forget 911 as an emergency service -- consider 411 > directory assistance, which is clearly NOT legislated. There is > agreement among the telcos what 411 will mean, and it isn't > legislated but it is a useful service goal. I would certainly like to see cooperation. However, the situation is such that cooperation is not possible. ICANN didn't cooperate with ORSC. ORSC has the resources, it seems, to deploy root servers with new TLDs. They believe they provide an acceptable level of service. There is no law (that I know of) that prevents them from doing this. After reading most of the arguments made between the factions on various mailing lists, I reached the conclusion fairly early out that there wasn't going to be any consensus, and that the matter would either be decided by individual choice (where a user decided to take their DNS service from) or by law (if an organization like the FCC were asked to step in and regulate DNS). > Fix ICANN, etc. Don't roll your own services unless you KNOW it > won't disrupt. See above. ORSC, etc. basically don't agree with ICANN-type models. > I believe the scope of this list suggests there is such a thing as a > netizen, or that network participation is participation in a > community. There is an unwritten obligation in any community not to > disrupt its infrastructure. The degree to which content served out of non-legacy root servers disrupts the infrastructure seems to be a widely debated topic where there is no consensus among concerned parties. That is another point that ORSC has argued. I can certainly envision some scenarios in which the load on all root servers would increase as a result of the existence of new TLDs in non-legacy root servers. I can certainly see where end users would be confused if they have gotten used to access to new TLDs in non-legacy root servers, then are forced to change ISPs (say, if their old ISP goes bankrupt) and the new ISP does not support the new TLDs. I don't doubt at all that these are problems; it is arguable whether they constitute substantial disruption to the infrastructure at large. A similar situation exists with ISPs who advertise routes with long prefixes. I don't know all the details, but some ISPs have entered into peering arrangements in order to propagate these prefixes, but others will only honor prefixes above a certain length (unless they fall into some particular category). This makes some parts of the Internet unreachable from others from an IP packet delivery standpoint. Again, while dealing with this type of problem can be irritating, it is debatable whether it constitutes significant disruption to the infrastructure. Now, as to the question of whether, say, new.net should compensate ISPs for any problems they incur with new root servers or new TLDs, before that question can be answered, more fundamental questions need to be answered, such as "What constitutes acceptable Internet service?" and "What are the obligations placed on the providers of Internet service?" I don't see how we could have a useful discussion on what, say, new.net "owes" any other organization, without resolving some of these more fundamental issues. > >Furthermore, there has never been a situation (that I know of) where > >911 service was funded through a competitive, virtually exclusive > >government contract, and then without warning or (imho) due process, > >the contract was amended to allow the providers to charge for the > >service. > I don't understand your analogy. In fact, enhanced 911 service is > often funded by legislated/utility commission imposed telephone usage > charges, putting the cost of 911 not on the overall community it > services, but on the subset of telephone users. I was actually making an analogy with NSI. NSI was granted a virtually exclusive contract to register domain names, and then granted the right to charge for the service at a time when it was most profitable to them. In effect, they were a monopoly, but under little if any regulation. 911 service, on the other hand, is the result of due process. There's also nothing particularly popular about 911 service that telephone companies should so greatly profit from it. > But that's just the point -- the present DNS works. It may not be > ideal from the standpoint of level playing fields, competitions, etc. > If people want level playing fields, put pressure on the US Justice > Department (there's no supranational body with jurisdiction) to take > antitrust action against NSI. Don't break the DNS or cause > operational disruption because people don't like ICANN, NSI, etc. How do you propose we put pressure on the US DoJ? The result of the pressure put on them was ICANN. My general feeling, as I said above, is that any result of US government intervention is going to be ICANN-like in nature. I haven't read anything from anyone in ICANN in about a year or so, but I don't recall any of them expressing any particular opinions about root servers or TLDs outside of their administrative scope. > I don't care if NSI does or does not have an unfair advantage, IF the > efforts to "level the playing field" damage innocent bystanders, Again, we come back to the question of to what extent non-universally resolvable domain names or extra load incurred on legacy root servers constitutes disruption of the infrastructure. > Well, try bringing up the idea that "useful names are a priority" in > an operational forum such as NANOG, and be sure you bring protective > clothing. The new.net proposal triggered a major uproar on the NANOG > mailing list, to the extent that the list sponsor ended the > discussion so other topics could be noticed. Go and check the NANOG > archives at www.nanog.org. > More than a "few people," and in a reasonably expert context. I don't read NANOG regularly, so I'll defer to your comments here. The people I've discussed these issues with do read NANOG, but do not participate in the debates regularly. Don't get me wrong. I don't agree entirely with everything ORSC endorses. Also, I am playing a bit of the devil's advocate here. But it is not clear to me that what new.net is doing is wrong, absent some consensus (which is not forthcoming) of what is right or even acceptable. - --gregbo ------------------------------ Date: Thu, 12 Apr 2001 17:19:50 -0700 (PDT) From: Greg Skinner Subject: Re: [netz] There is a need for online discussion of new DNS NAS Commmittee (hcb intended to send this message to the list. I'm resending it on his behalf.) >hcb wrote: >> gds wrote: >> >I don't understand the point you're trying to make here. Are you >> >trying to argue that what new.net is doing is requiring other parties >> >(e.g. ISPs) to do something, but new.net is not funding them to do it? >> Exactly. Their profit making activity incurs costs on the part of >> others who do not share in those profits. > >Could you clarify what you mean by "requiring ISPs to do something?" >What specifically are ISPs required to do, as a result of what new.net >is doing? New.net has declared an assortment of new top-level domains, including .shop, .mp3, and .chat. I am a nontechnical user, whose access does not come from one of the providers with whom new.net contracts. I see an ad for www.foo.shop, and type the URL into Netscape. Netscape returns an error message that it cannot find the domain. I doublecheck the ad in which I saw www.foo.shop, and then decide something is broken. I call tech support at my ISP. My ISP runs a standard DNS. One option is that they tell their customer that this is a domain name that isn't part of the generally accepted domain name system. Another option is for the ISP to contract with new.net. Yet another option is to help the user get a browser plugin that supports new.net domains. It's also very realistic that the first-level support people at the ISP will have no idea what is causing the problem, and may escalate the call to the DNS administrator of the ISP -- who already is busy. But when the ISP budgeted its support functions, it at least informally was based on experience with the number of DNS-related support calls it would need to handle. There is now a whole new class of problems to deal with, an increased workload, etc. If the ISP incurs cost to be responsive, what benefit is there to the ISP? Customer retention, I suppose. But it gives me a nasty reminder of a Mob protection racket -- contract with us for protection/special DNS and there won't be any accidents to your customers. > >> >I don't think useful comparisons can be made between DNS service (as >> >it is currently practiced) and 911 service. [...] >> There is a useful comparison, if you accept the idea of cooperation >> rather than anyone being free to do their own thing. I can have >> confidence that if I dial 911, I will not be connected to a pizza >> delivery service. Forget 911 as an emergency service -- consider 411 >> directory assistance, which is clearly NOT legislated. There is >> agreement among the telcos what 411 will mean, and it isn't >> legislated but it is a useful service goal. > >I would certainly like to see cooperation. However, the situation is >such that cooperation is not possible. ICANN didn't cooperate with >ORSC. ORSC has the resources, it seems, to deploy root servers with >new TLDs. They believe they provide an acceptable level of service. >There is no law (that I know of) that prevents them from doing this. There's no law authorizing ICANN to do anything. What jurisdiction would apply? > >After reading most of the arguments made between the factions on >various mailing lists, I reached the conclusion fairly early out that >there wasn't going to be any consensus, and that the matter would >either be decided by individual choice (where a user decided to take >their DNS service from) or by law (if an organization like the FCC were >asked to step in and regulate DNS). > >> Fix ICANN, etc. Don't roll your own services unless you KNOW it >> won't disrupt. > >See above. ORSC, etc. basically don't agree with ICANN-type models. Fine. I don't agree with ORSC models. I will support ICANN, flawed as it may be, because it doesn't disrupt that which will work. As an invididual, I will oppose any self-appointed "stakeholder" that wants to play its own DNS games. I rely on the IAB and IETF as definitive as what is supported in the general Internet. > >> I believe the scope of this list suggests there is such a thing as a > > netizen, or that network participation is participation in a >> community. There is an unwritten obligation in any community not to >> disrupt its infrastructure. > >The degree to which content served out of non-legacy root servers >disrupts the infrastructure seems to be a widely debated topic where >there is no consensus among concerned parties. Depends on what you mean by "concerned parties." Ronda goes farther than I do in assuming the traditional scientific community that created the Internet are the "good guys." > That is another point >that ORSC has argued. I can certainly envision some scenarios in >which the load on all root servers would increase as a result of the >existence of new TLDs in non-legacy root servers. I can certainly see >where end users would be confused if they have gotten used to access >to new TLDs in non-legacy root servers, then are forced to change ISPs >(say, if their old ISP goes bankrupt) and the new ISP does not support >the new TLDs. I don't doubt at all that these are problems; it is >arguable whether they constitute substantial disruption to the >infrastructure at large. > >A similar situation exists with ISPs who advertise routes with long >prefixes. I don't know all the details, but some ISPs have entered >into peering arrangements in order to propagate these prefixes, but >others will only honor prefixes above a certain length (unless they >fall into some particular category). This makes some parts of the >Internet unreachable from others from an IP packet delivery >standpoint. Again, while dealing with this type of problem can be >irritating, it is debatable whether it constitutes significant >disruption to the infrastructure. Excuse me. You are getting into my own area of specialization, which is global Internet routing, and where I think I can legitimately say I do know the relevant details. It's much more complex than you describe, or can be explained in a few paragraphs. As you say, "only honor prefixes that fall into some category." Lots and lots of categories. With current projections, there will be major breakages in the IPv4 routing system within the moderate term -- 2004 to 2008. I am actively involved in work to develop short- to moderate-term fixes to extend these limits, which involve both operational practices and new design approaches for routers. The Internet Research Task Force has been chartered to look at a longer-term new routing architecture. Incidentally, IPv6 isn't a panacea here. The existing BGP interdomain routing architecture is being asked to do things it was never designed to do, such as extensive enterprise multihoming and traffic engineering. For that matter, DNS is being asked to do things it was never designed to do, such as be a general-purpose, user-friendly resource location system. In the case of IP routing, the operations community is far more cooperative than the DNS crew -- there would be no question, for example, of any support of anyone that tried to announce address space that a "grass roots registry" decided to assign to them. Please be more specific about what you are saying is or is not debatable. I can't quite understand the point--what you thing is good and bad. > >Now, as to the question of whether, say, new.net should compensate >ISPs for any problems they incur with new root servers or new TLDs, >before that question can be answered, more fundamental questions need >to be answered, such as "What constitutes acceptable Internet >service?" and "What are the obligations placed on the providers of >Internet service?" I don't see how we could have a useful discussion >on what, say, new.net "owes" any other organization, without resolving >some of these more fundamental issues. > >> >Furthermore, there has never been a situation (that I know of) where >> >911 service was funded through a competitive, virtually exclusive >> >government contract, and then without warning or (imho) due process, >> >the contract was amended to allow the providers to charge for the >> >service. >> I don't understand your analogy. In fact, enhanced 911 service is >> often funded by legislated/utility commission imposed telephone usage > > charges, putting the cost of 911 not on the overall community it >> services, but on the subset of telephone users. > >I was actually making an analogy with NSI. NSI was granted a >virtually exclusive contract to register domain names, and then >granted the right to charge for the service at a time when it was most >profitable to them. In effect, they were a monopoly, but under little >if any regulation. 911 service, on the other hand, is the result of >due process. There's also nothing particularly popular about 911 >service that telephone companies should so greatly profit from it. > >> But that's just the point -- the present DNS works. It may not be >> ideal from the standpoint of level playing fields, competitions, etc. >> If people want level playing fields, put pressure on the US Justice >> Department (there's no supranational body with jurisdiction) to take >> antitrust action against NSI. Don't break the DNS or cause >> operational disruption because people don't like ICANN, NSI, etc. > >How do you propose we put pressure on the US DoJ? The result of the >pressure put on them was ICANN. My general feeling, as I said above, >is that any result of US government intervention is going to be >ICANN-like in nature. The congressional appropriations committee for the US DoJ is a very good start. For that matter, a direct complaint to the Antitrust Division of the DoJ. I'm not going to do that, because I don't consider the situation sufficiently broken to care about. > >I haven't read anything from anyone in ICANN in about a year or so, >but I don't recall any of them expressing any particular opinions >about root servers or TLDs outside of their administrative scope. > >> I don't care if NSI does or does not have an unfair advantage, IF the >> efforts to "level the playing field" damage innocent bystanders, > >Again, we come back to the question of to what extent non-universally >resolvable domain names or extra load incurred on legacy root servers >constitutes disruption of the infrastructure. > >> Well, try bringing up the idea that "useful names are a priority" in >> an operational forum such as NANOG, and be sure you bring protective >> clothing. The new.net proposal triggered a major uproar on the NANOG >> mailing list, to the extent that the list sponsor ended the >> discussion so other topics could be noticed. Go and check the NANOG >> archives at www.nanog.org. > >> More than a "few people," and in a reasonably expert context. > >I don't read NANOG regularly, so I'll defer to your comments here. >The people I've discussed these issues with do read NANOG, but do not >participate in the debates regularly. I'm a long-time participant in NANOG. I can't deal with vague comments about "people who read but don't participate." > >Don't get me wrong. I don't agree entirely with everything ORSC >endorses. Also, I am playing a bit of the devil's advocate here. >But it is not clear to me that what new.net is doing is wrong, absent >some consensus (which is not forthcoming) of what is right or even >acceptable. > >--gregbo ------------------------------ End of Netizens-Digest V1 #381 ******************************