Heheh.. :) Speakeasy if I recall correctly, just got started in IPv6 core development at their backbone, but does not have v6 transit yet. They peer with a few networks, like C&W, etc (occaid will be peering with them also shortly, just need to figure out what exchange point IPs are being used in Atlanta).
The "immeasurably inferior[tm]" routing path you are seeing is most likely a result of Rogue ASes who leak peer-learned routes to non-customers, such as legacy tunnels, which we run into an issue like this :( Unfortunately, HE.net does play a significant role in these issues as they have no sensible IPv6 routing policy. This issue also happened to DTAG/Deutsche Telekom's brand new /19 IPv6 space. At that time, they did not acquire transit yet, and Rogue ASes were already leaking their peering relationships when they should not be, over tunnels. Regards, James Jun IP Infrastructure & Technology Services TowardEX Technologies, Inc. WWW: http://www.towardex.com Email: [EMAIL PROTECTED] Office: +1 (617) 459-4051 Ext. 179 Mobile: +1 (978) 394-2867 > -----Original Message----- > From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of > Jeroen Massar > Sent: Tuesday, January 03, 2006 4:30 PM > To: Kenneth Porter > Cc: users@ipv6.org > Subject: Re: ipv6 dns server. > > Kenneth Porter wrote: > > --On Monday, January 02, 2006 4:02 AM +0800 Lawrence Hughes > > <[EMAIL PROTECTED]> wrote: > > > >> Most client computers (DNS resolvers) that support IPv6 will (and > should) > >> use the IPv6 addresses preferentially over IPv4 when both are returned > >> from the DNS. > > > > With most ISP's not providing an IPv6 gateway, is that yet wise? > > AFAIK Windows is currently the only OS that doesn't resolve using IPv6 > as a transport. Most *ix (*BSD/Linux) implementations do. > > Also note that if a DNS server is configured to use both IPv4 and IPv6 > as a transport it will first try IPv6 to contact the DNS server in > question and after that IPv4. > > In most setups I have encountered there was a dual-stack DNS server > which would speak to other DNS servers using IPv6 where possible. It > properly falls back to IPv4 when noted that the IPv6 server is > unreachable or gives slow responses, those are default properties of the > DNS protocol. > > I have not yet heard any complaints from folks who where using these > kind of setups yet. So it appears to work pretty well. > > Notez bien there are no published IPv6 root-servers and one will need a > dual-stack DNS server somewhere to be able to reach about 99% of the > Internet anyway. For an endhost, using only IPv6 as a DNS transport is > of course very well possible and should not cause any problems, unless > your connectivity to the DNS server goes down ;) > > > Even Speakeasy, one of the more technically competent ISP's, doesn't yet > > provide native IPv6. > > From GRH (http://www.sixxs.net/tools/grh/dfp/arin/): > 2001:1858::/32 [us] SPEK-V6-0 Speakeasy Network > Allocated: 2003-08-07 > First Announced: 2005-12-07 13:33:56 > Last seen: 2006-01-03 22:17:22 > > Apparently they have connectivity, at least the BGP route is there, it > is actually severely broken due to "HE.net's immeasurably superior IPv6 > routing" (read: they are playing Tier-1 without being one), see below. > > Also it is very simple to solve, if they don't provide it: tunnel it! > Check: http://en.wikipedia.org/wiki/Tunnel_Broker > > Tunneling of course should only be done when one is an endsite, tunnels > should not be used for transit also see: > http://ip6.de.easynet.net/ipv6-minimum-peering.txt > > Greets, > Jeroen > > -- > traceroute to 2001:1858::1 (2001:1858::1) from > 2001:7b8:20d:0:20c:29ff:fe36:4f, 30 hops max, 16 byte packets > 1 purgatory.unfix.org (2001:7b8:20d:0:290:27ff:fe24:c19f) 25.292 ms > 8.73 ms 8.042 ms > 2 2001:7b8:5:10:74::1 (2001:7b8:5:10:74::1) 12.208 ms 21.739 ms > 13.612 ms > 3 i49.ge-0-1-0.jun1.kelvin.ipv6.network.bit.nl > (2001:7b8:3:31:290:6900:31c6:d81f) 17.384 ms 13.5 ms 19.189 ms > 4 jun1.sara.ipv6.network.bit.nl (2001:7b8::205:8500:120:7c1f) 13.987 > ms 6.541 ms 13.334 ms > 5 v6-transit.glbx.net (2001:7b8:40:7::1) 4.856 ms 7.186 ms 9.46 ms > 6 eth10-0-0.xr1.ams1.gblx.net (2001:7f8:1::a500:3549:1) 12.342 ms > 11.113 ms 10.055 ms > 7 nl-ams04a-re1-fe-0-0.ipv6.aorta.net (2001:7f8:1::a500:6830:1) > 11.989 ms 10.971 ms 9.904 ms > 8 nl-ams06d-re1-t-2.ipv6.aorta.net (2001:730::1:c) 9.937 ms 16.968 > ms 11.934 ms > 9 hurrican.net-gw1.nl.ipv6.aorta.net (2001:730::1:2f) 102 ms 109.078 > ms 113.883 ms > 10 3ffe:81d0:ffff:1::1 (3ffe:81d0:ffff:1::1) 128.097 ms 119.002 ms > 109.196 ms > 11 3ffe:80a::e (3ffe:80a::e) 121.231 ms 133.967 ms 132.948 ms > 12 * * * > 13 * * * > 14 * * > > Hop 9 is Hurricane Electric > Hop 10 is Hurricane Electric's 6bone address space (going away 6/6/6) > Hop 11 is ISI-LAP > then it gets lost in 6bone space... > --------------------------------------------------------------------- The IPv6 Users Mailing List Unsubscribe by sending "unsubscribe users" to [EMAIL PROTECTED]
