ap@conference.ripe.net/2006-04-26.txt:[13:04:58] I am Alex Le Heux from the RIPE NCC and I will monitor this session. If you have any questions I can read them out for you. ap@conference.ripe.net/2006-04-26.txt:[13:05:24] Please state your name as well when you want me to ask a question. ap@conference.ripe.net/2006-04-26.txt:[13:06:09] * nick has joined the channel. ap@conference.ripe.net/2006-04-26.txt:[13:06:11] * davew has joined the channel. ap@conference.ripe.net/2006-04-26.txt:[13:07:42] * colin has joined the channel. ap@conference.ripe.net/2006-04-26.txt:[13:09:56] --- arno has become available ap@conference.ripe.net/2006-04-26.txt:[13:11:37] * brett has joined the channel. ap@conference.ripe.net/2006-04-26.txt:[13:12:10] For the late arrivals in this chatroom: I'm Alex from the RIPE NCC and will monitor this session. If you have any questions I can read them out for you. Please state your full name when you do want me to ask something. ap@conference.ripe.net/2006-04-26.txt:[13:14:23] --- juampe has become available ap@conference.ripe.net/2006-04-26.txt:[13:15:49] --- rhe has become available ap@conference.ripe.net/2006-04-26.txt:[13:16:42] --- rumy has become available ap@conference.ripe.net/2006-04-26.txt:[13:19:00] * Niels has joined the channel. ap@conference.ripe.net/2006-04-26.txt:[13:19:10] --- Niels has left ap@conference.ripe.net/2006-04-26.txt:[13:20:08] --- webmaster-ncc has become available ap@conference.ripe.net/2006-04-26.txt:[13:22:38] --- adrian has become available ap@conference.ripe.net/2006-04-26.txt:[13:23:16] --- ballroom has become available ap@conference.ripe.net/2006-04-26.txt:[13:25:15] --- caz has become available ap@conference.ripe.net/2006-04-26.txt:[13:27:21] --- cfriacas has become available ap@conference.ripe.net/2006-04-26.txt:[13:30:47] --- pheldal has become available ap@conference.ripe.net/2006-04-26.txt:[13:34:36] * Niels quit IRC altogether ap@conference.ripe.net/2006-04-26.txt:[13:36:35] * Niels has joined the channel. ap@conference.ripe.net/2006-04-26.txt:[13:40:27] * mike_5459 has joined the channel. ap@conference.ripe.net/2006-04-26.txt:[13:41:42] * master has joined the channel. ap@conference.ripe.net/2006-04-26.txt:[13:42:14] --- jhma has become available ap@conference.ripe.net/2006-04-26.txt:[13:51:38] * Remco-16243 has joined the channel. ap@conference.ripe.net/2006-04-26.txt:[13:51:54] (Remco-16243) can somebody please turn down the volume of our esteemed speaker :) ap@conference.ripe.net/2006-04-26.txt:[13:51:59] (Niels) how does one not qualify for becoming an LIR? ap@conference.ripe.net/2006-04-26.txt:[13:52:16] (master) not having "customers" to assign address space to? ap@conference.ripe.net/2006-04-26.txt:[13:52:22] if one aims to save some euros/year :-) ap@conference.ripe.net/2006-04-26.txt:[13:52:30] (Djinh) anyone can qualify ap@conference.ripe.net/2006-04-26.txt:[13:52:31] Could be that they are not in the RIPE NCC service region ap@conference.ripe.net/2006-04-26.txt:[13:52:34] (Djinh) customers are not a requirement ap@conference.ripe.net/2006-04-26.txt:[13:52:38] (Remco-16243) that's what the 'enterprise' lir is for isn't it ? ap@conference.ripe.net/2006-04-26.txt:[13:52:41] (Niels) enterprise LIR doesn't have customers ap@conference.ripe.net/2006-04-26.txt:[13:52:54] (nick) from a lir point of view, they do ap@conference.ripe.net/2006-04-26.txt:[13:52:59] (Djinh) there is no longer a distinction between isp and enterprise lirs ap@conference.ripe.net/2006-04-26.txt:[13:53:03] (Niels) webmaster - that's plausible but then would they qualify for a PI v6 assignment in the first place? ap@conference.ripe.net/2006-04-26.txt:[13:53:07] (master) yes, i rather spend my money on more uplinks and better routers than for a LIR contract if i don't really need to assign address space ap@conference.ripe.net/2006-04-26.txt:[13:53:49] If anyone has any questions for the microphone, please let me know. ap@conference.ripe.net/2006-04-26.txt:[13:54:48] Niels, no. If they do not have a registered office in the NCC Service region, they cannot become a member of the RIPE NCC. And if they are not a member, they cannot get any resources directly from the RIPE NCC. ap@conference.ripe.net/2006-04-26.txt:[13:55:37] yes, but that is the service area issue... ap@conference.ripe.net/2006-04-26.txt:[13:55:43] (Djinh) you do not need a registered office in the ripe region to qualify for an IPv4 PI assignment, just the intention to use it in the ripe region is enough. ap@conference.ripe.net/2006-04-26.txt:[13:55:46] Indeed ap@conference.ripe.net/2006-04-26.txt:[13:56:12] (Niels) ok, so the "(if qualify [sic[ for it)" caveat was probably meant that way, then. thanks ap@conference.ripe.net/2006-04-26.txt:[13:58:29] * tmind has joined the channel. ap@conference.ripe.net/2006-04-26.txt:[13:59:05] (master) *Sigh* the whole discussion always was ridiculous... everyone can get LIR, there is Extra-Small now, so this practically PI with a recurring Price-Tag on it ap@conference.ripe.net/2006-04-26.txt:[13:59:25] (master) won't save anything in the routing table in the end ap@conference.ripe.net/2006-04-26.txt:[13:59:28] yep ap@conference.ripe.net/2006-04-26.txt:[13:59:30] (Niels) except that if you don't have 200 sites you won't get v6 PA space, even as an LIR ap@conference.ripe.net/2006-04-26.txt:[13:59:44] (tmind) afrinic moved that route voor v4 PI I noticed (recurring fee for PI owners) ap@conference.ripe.net/2006-04-26.txt:[13:59:45] (master) that's gonna be scrapped anyways soon :) ap@conference.ripe.net/2006-04-26.txt:[14:00:01] Niels: That's covered by 2005-03 (I think). ap@conference.ripe.net/2006-04-26.txt:[14:00:08] (master) jup ap@conference.ripe.net/2006-04-26.txt:[14:00:20] i think the 200-rule is going to be removed sooner or later... ap@conference.ripe.net/2006-04-26.txt:[14:01:05] (master) it's of no use either see Mailing list discussions... anyone really got 200 customers by now? ap@conference.ripe.net/2006-04-26.txt:[14:01:21] (master) s/200 customers/200 Assignments/ ap@conference.ripe.net/2006-04-26.txt:[14:03:42] (Niels) I'm sure XS4ALL has made a few thousand tunnel assignments for its customers, for example ap@conference.ripe.net/2006-04-26.txt:[14:04:09] (Niels) and each colo'd customer box they have has a /48, doesn't it (pinging Djinh)? ap@conference.ripe.net/2006-04-26.txt:[14:04:10] (master) Tunnel... well OK does that count as "customer"? good question ap@conference.ripe.net/2006-04-26.txt:[14:04:23] (Djinh) niels, yes that's true ap@conference.ripe.net/2006-04-26.txt:[14:04:35] (Niels) they only offer tunnels to customers, one tunnel per customer, so I guess it counts as a site, why wouldn't it? ap@conference.ripe.net/2006-04-26.txt:[14:04:43] (Djinh) master, it counts as a valid 'assignment' in every sense of the word ap@conference.ripe.net/2006-04-26.txt:[14:04:50] (Niels) (I'm a customer; I get one address on the /64 from the tunnel, and a /48 routed down it.) ap@conference.ripe.net/2006-04-26.txt:[14:05:14] (master) niels: only to customers - ok then didn't know that, "tunnel" usually is to "anyone who wants it" for me :) ap@conference.ripe.net/2006-04-26.txt:[14:05:33] (master) djinh: right, but i don't think that's the original intention of the 200-customer rule, but i want it gone anyways :) ap@conference.ripe.net/2006-04-26.txt:[14:05:39] (Niels) XS4ALL is smarter than average ;-) ap@conference.ripe.net/2006-04-26.txt:[14:06:20] Wrt the fear of re-numbering; a criteria for a furure decision to set an expiry date must be a requirement for tools that eliminate the pain of renumbering. Any future so-called scaleable solution is DOA unless it comes with such tools. ap@conference.ripe.net/2006-04-26.txt:[14:06:25] (Djinh) master, the 200 rule definitely applies to people like DSL customers ap@conference.ripe.net/2006-04-26.txt:[14:06:38] --- webmaster-ncc has left ap@conference.ripe.net/2006-04-26.txt:[14:07:47] (master) djinh: sure, that are your customers, but "tunnel" is usually used from tunnel-brokers who give out a /48 to just anyone requesting it all over the world. I consider Tunnel-assignments for actual customers as perfectly couting, but not Tunnels to just anyone ap@conference.ripe.net/2006-04-26.txt:[14:07:56] pheldal, dpo you want me to relay that to the microphone? ap@conference.ripe.net/2006-04-26.txt:[14:08:53] (nick) leo's proposal requires that RIPE charge for all PI address space, not just ipv6 ap@conference.ripe.net/2006-04-26.txt:[14:09:00] (nick) this is a Good Thing ap@conference.ripe.net/2006-04-26.txt:[14:09:13] (master) all other RIRs do that i think ap@conference.ripe.net/2006-04-26.txt:[14:09:24] (nick) throwing out PI space without having any means to reclaim it is poor operational practice ap@conference.ripe.net/2006-04-26.txt:[14:09:27] (Niels) nick - agree ap@conference.ripe.net/2006-04-26.txt:[14:09:56] (master) Actually i would prefer a one-time fee for PI in contrast to recurring fee as LIR, but that doesn't solve the problem of orphaned PI Space so it's not such a good idea :-( ap@conference.ripe.net/2006-04-26.txt:[14:10:25] (nick) however, charging for PI ipv6 space will create another barrier to deployment ap@conference.ripe.net/2006-04-26.txt:[14:10:40] (nick) barriers are bad, considering the especially poor uptake of ipv6 space ap@conference.ripe.net/2006-04-26.txt:[14:10:41] (master) tjum ap@conference.ripe.net/2006-04-26.txt:[14:10:41] (Remco-16243) if we're going to do PI we're going to have a swamp in ipv6 anyway ap@conference.ripe.net/2006-04-26.txt:[14:10:47] (Remco-16243) so why reclaim stuff that's in there ap@conference.ripe.net/2006-04-26.txt:[14:11:09] (master) IPv6 won't be widly deployed until IPv4 runs out of address space ... so no price tag for Ipv4 PI and no more request forms to make it run out faster? :) ap@conference.ripe.net/2006-04-26.txt:[14:11:29] I made a proposal? ap@conference.ripe.net/2006-04-26.txt:[14:11:44] (nick) remco: let's create a swamp then. but with the realisation that if you give each organisation a large block, that will be all they will ever need for their ipvnetwork requirements ap@conference.ripe.net/2006-04-26.txt:[14:11:56] (nick) assigning pi space implies a swamp ap@conference.ripe.net/2006-04-26.txt:[14:12:09] (Remco-16243) that's my point ap@conference.ripe.net/2006-04-26.txt:[14:12:42] (nick) ergo the ipv6 swamp will be smaller than the v4 swamp ap@conference.ripe.net/2006-04-26.txt:[14:12:48] (tmind) current pi proposal talks about /32's, so what's the difference between a lir or pi route :) ap@conference.ripe.net/2006-04-26.txt:[14:13:12] (master) the ARIN proposal says /48 AFAICR ap@conference.ripe.net/2006-04-26.txt:[14:13:13] (Remco-16243) your address is in a bad neighbourhood ? :) ap@conference.ripe.net/2006-04-26.txt:[14:13:26] (mike_5459) ...and we're assuming the 200 number is going? ap@conference.ripe.net/2006-04-26.txt:[14:13:30] (master) The problem with anythng but /32 and /35 is the current filter practice ap@conference.ripe.net/2006-04-26.txt:[14:13:34] (tmind) (and indeed, what will as such be the difference between just becoming a lir + v6 alloc, assuming the 200 assign rules disappears) and PI? ap@conference.ripe.net/2006-04-26.txt:[14:13:43] If anyone wants me to relay something to the microphone, please let me know. If you would then also include your full name, that would make our scribes very happy :) ap@conference.ripe.net/2006-04-26.txt:[14:14:02] (mike_5459) or we have a bunch of people who dissapear down the crack between not being an end site, yet not making 200 assignments? ap@conference.ripe.net/2006-04-26.txt:[14:14:03] (nick) /32 is simply too large, and removes any justification to become a LIR ap@conference.ripe.net/2006-04-26.txt:[14:14:04] (master) tmind: that as LIR you probably pay for things you don't need as non-assigning End-customer ap@conference.ripe.net/2006-04-26.txt:[14:14:08] PF-len for PI doesn't really matter. It's the size of the RIB that people care about. ap@conference.ripe.net/2006-04-26.txt:[14:14:10] (Remco-16243) i don't consider filtering a problem right now as the actual real world ipv6 routing table seems to be shrinking at the moment ap@conference.ripe.net/2006-04-26.txt:[14:14:10] I'd rather fix filters than burn huge amounts of address space... ap@conference.ripe.net/2006-04-26.txt:[14:14:13] --- caz has left: Disconnected ap@conference.ripe.net/2006-04-26.txt:[14:14:28] (Niels) nick, anything smaller will make you already run into filters, but yes ap@conference.ripe.net/2006-04-26.txt:[14:14:35] (nick) i mean, a /48 is 16384 networks ap@conference.ripe.net/2006-04-26.txt:[14:15:03] (Remco-16243) how many PI allocations are out there in ipv4 space ? ap@conference.ripe.net/2006-04-26.txt:[14:15:05] (tmind) master: oki, so we get the point, its the difference between a one-off cost, paid by your ISP because you want PI, or paying for it yourself because you want to be part of the DFZ? ap@conference.ripe.net/2006-04-26.txt:[14:15:06] (nick) that's a very, very big network deployment ap@conference.ripe.net/2006-04-26.txt:[14:15:12] (master) remco: maybe but look at the problems with new IPv4 /8 Allocations from reserved space... the pain to get all people to update their filters.... some won't update their /32 Filters ap@conference.ripe.net/2006-04-26.txt:[14:15:58] (master) tmind: as LIR you have to do all the Assignment stuff on your own, have more responsibility probably and such stuff you might not want ap@conference.ripe.net/2006-04-26.txt:[14:16:01] (Remco-16243) master, i think the amount of unmaintained ipv6 routers is significantly less than the amount of unmaintained ipv4 routers ap@conference.ripe.net/2006-04-26.txt:[14:16:07] (Niels) Remco - I have a private interconnect with a party at work, and they won't accept my /48 announcements over that ap@conference.ripe.net/2006-04-26.txt:[14:16:13] (tmind) (what would be the the numbers of points one gets with a single v6 alloc +asn? small/extra-small?) ap@conference.ripe.net/2006-04-26.txt:[14:16:17] (master) i mean, i had customers complain about LIR mails and announcements "we don't want that" and so ap@conference.ripe.net/2006-04-26.txt:[14:16:26] (Remco-16243) niels, in that case, they suck :) ap@conference.ripe.net/2006-04-26.txt:[14:16:32] (master) remco: yes but the amount of hard-core Filter advocates is higher in Ipv6 :) ap@conference.ripe.net/2006-04-26.txt:[14:17:02] (Remco-16243) so they'll cut them and their customers off from significant portions of ipv6 ap@conference.ripe.net/2006-04-26.txt:[14:17:05] (Remco-16243) their choice :) ap@conference.ripe.net/2006-04-26.txt:[14:17:06] (nick) what's wrong with a /40 assignment? that's 16E6 networks ap@conference.ripe.net/2006-04-26.txt:[14:17:19] The "hard-core filter advocates" aren't the ones to worry about -- they'll be quite responsive at updating their filters. ap@conference.ripe.net/2006-04-26.txt:[14:17:34] (master) well i announce some /40s ... virtually no but some problems in reachability when i drop the /32 ... but better than /48 announcements, yes ap@conference.ripe.net/2006-04-26.txt:[14:17:39] (tmind) master: if they don't want that they will for sure not want to care about the whole discussion we have about renumbering etc here, it's just 'give me the addresses and asn I pay for and stop buggering me'? ap@conference.ripe.net/2006-04-26.txt:[14:17:59] (master) rhe: i don't really think so but not sure about that either, just a feeling ap@conference.ripe.net/2006-04-26.txt:[14:18:18] (master) tmind: something like that, yes ap@conference.ripe.net/2006-04-26.txt:[14:18:29] google.com ap@conference.ripe.net/2006-04-26.txt:[14:18:57] (master) The ones with money always will get "PI" the one way or the other (getting LIR and not caring about any LIR duties) ap@conference.ripe.net/2006-04-26.txt:[14:19:10] (Remco-16243) exactly ap@conference.ripe.net/2006-04-26.txt:[14:19:19] --- cfriacas has left ap@conference.ripe.net/2006-04-26.txt:[14:19:22] (master) i don't really think unresponsive and !clueful LIRs are the way to go ap@conference.ripe.net/2006-04-26.txt:[14:19:30] (Remco-16243) 3500euro a year is nothing for a large enterprise ap@conference.ripe.net/2006-04-26.txt:[14:19:39] (tmind) master: so, why would such a party ever want to renumber afterwards, it will be a kind of 'permanent' /32 that way, why make the difference and not make them lir and pay for it? ap@conference.ripe.net/2006-04-26.txt:[14:19:46] (nick) but prohibitive for a small operation with small needs ap@conference.ripe.net/2006-04-26.txt:[14:20:22] --- cfriacas has become available ap@conference.ripe.net/2006-04-26.txt:[14:20:36] * will has joined the channel. ap@conference.ripe.net/2006-04-26.txt:[14:20:59] (master) tmind: that's the "i don't care about money" fraction only, they cause more problems than do good i think. I personally rather care about the people with clue but no money, non-commercial organisations ect. Them not being allowed in the global table just because they are "not big enough" is a bit... problematic to me ap@conference.ripe.net/2006-04-26.txt:[14:21:11] If v6 is ready for full deployment, then there is currently no viable alternative to some form of PI. I would however urge people to re-think because; (1) v6 at this point doesn't add significant functionality at the application layers. (2) v6 without scaleable routing solutions is at best described as unfinished. (3) If we still consider v6 experimental can't we live with allocation policies that don't accomodate every possible need? ap@conference.ripe.net/2006-04-26.txt:[14:21:16] (master) on the other hand i see the Orphan-problem without a recurring cost, too :-/ ap@conference.ripe.net/2006-04-26.txt:[14:21:36] --- caz has become available ap@conference.ripe.net/2006-04-26.txt:[14:22:13] pheldal, if you want me to relay that to the mike, can you state your full name please? ap@conference.ripe.net/2006-04-26.txt:[14:22:58] --- ferenc_c has become available ap@conference.ripe.net/2006-04-26.txt:[14:23:34] http://www.nro.net/documents/nro34.html#3-1 ap@conference.ripe.net/2006-04-26.txt:[14:24:22] --- rhe has left: Logged out ap@conference.ripe.net/2006-04-26.txt:[14:24:27] If v6 is to be considered incomplete it would be nice if influential people in the IETF-environment would communicate that loud and clear, so that the .gov sector among others don't turn v6 into another GOSIP ap@conference.ripe.net/2006-04-26.txt:[14:24:44] --- rhe has become available ap@conference.ripe.net/2006-04-26.txt:[14:25:28] Djinh: not really necessary. aren't irc/jabbel logs supposed to be archived? ap@conference.ripe.net/2006-04-26.txt:[14:25:52] Yes, it is archived ap@conference.ripe.net/2006-04-26.txt:[14:26:07] (master) No gert, i wasn't shouding at you, i was applauding :) ap@conference.ripe.net/2006-04-26.txt:[14:26:40] --- jhma has left ap@conference.ripe.net/2006-04-26.txt:[14:26:48] --- adrian has left ap@conference.ripe.net/2006-04-26.txt:[14:27:08] --- ferenc_c has left ap@conference.ripe.net/2006-04-26.txt:[14:27:23] This session has now ended, goodbye!