From andy at nosignal.org Tue Aug 3 23:57:45 2010 From: andy at nosignal.org (Andy Davidson) Date: Tue, 3 Aug 2010 22:57:45 +0100 Subject: Cisco 1801W, v6 on wireless, bvi argh. In-Reply-To: <4BF2672D.9080508@netability.ie> References: <24D585B4-F4CD-4CC6-AE98-46C1A3BEE4C7@nosignal.org> <4BF0538D.2000005@netability.ie> <4BF26150.9030407@netability.ie> <8DB354A9-C161-461C-A787-EB7C8EA40E2C@employees.org> <4BF2672D.9080508@netability.ie> Message-ID: On 18 May 2010, at 11:08, Nick Hilliard wrote: > On 18/05/2010 10:58, Ole Troan wrote: >>>> You didn't try 12.4(22)T, did you? This bug was fixed in that version, and >>> >>> I lied - the bug wasn't fixed in that version after all. >>> >>> I could have sworn at the time that either this bugid or a similar bugid >>> was marked fixed in 12.4(22)T. However, looking at the description for >>> CSCej50923, there's no fixed listed for 12.4T at all. Same with bugid >>> CSCta27529, which was mentioned in: >> >> CSCta27529 is integrated in 15.1T images. > > Both CSCej50923 and CSCta27529 are listed as being fixed in 15.1(0.1)T > onwards. However, if as Andy claims there is no "bridge XXX route ipv6" > command in 15.1(1)T, then it looks like there's something odd going on. Well, I can't explain the following based on the information in the bug tool, but I upgraded to 15.1(2)T (Adv Enterprise), and ipv6 is working on the bvi interface right now. bridge XXX router ipv6 still does not appear as a possible command, but I put the ipv6 address on the BVI interface, and everything is singing and dancing as planned. *****Thank you***** to anyone on list who was involved with making this work. Bits of config that might serve as a hint, get in touch offlist for more ... ipv6 unicast-routing ipv6 cef ipv6 multicast-routing ! bridge irb ! interface Dot11Radio0 description Wireless Network / vlan 2 no ip address ! ! interface Dot11Radio0.2 encapsulation dot1Q 2 bridge-group 1 bridge-group 1 subscriber-loop-control bridge-group 1 spanning-disabled bridge-group 1 block-unknown-source no bridge-group 1 source-learning no bridge-group 1 unicast-flooding ! interface BVI1 description Wireless network VLAN ipv6 address xxxx:xxxx:100:2::1/64 ipv6 enable ! bridge 1 protocol ieee bridge 1 route ip -- Regards, Andy Davidson +44 (0)20 7993 1700 www.netsumo.com NetSumo Specialist networks consultancy for ISPs, Whitelabel 24/7 NOC From gert at space.net Wed Aug 4 11:54:49 2010 From: gert at space.net (Gert Doering) Date: Wed, 4 Aug 2010 11:54:49 +0200 Subject: Cisco 1801W, v6 on wireless, bvi argh. In-Reply-To: References: <24D585B4-F4CD-4CC6-AE98-46C1A3BEE4C7@nosignal.org> <4BF0538D.2000005@netability.ie> <4BF26150.9030407@netability.ie> <8DB354A9-C161-461C-A787-EB7C8EA40E2C@employees.org> <4BF2672D.9080508@netability.ie> Message-ID: <20100804095449.GK73473@Space.Net> Hi, On Tue, Aug 03, 2010 at 10:57:45PM +0100, Andy Davidson wrote: > Well, I can't explain the following based on the information in > the bug tool, but I upgraded to 15.1(2)T (Adv Enterprise), and ipv6 > is working on the bvi interface right now. > > bridge XXX router ipv6 still does not appear as a possible command, > but I put the ipv6 address on the BVI interface, and everything is > singing and dancing as planned. At last :-) - Thanks for this update. > Bits of config that might serve as a hint, get in touch offlist for more ... If this box isn't in production use, could you test something? > interface Dot11Radio0 > description Wireless Network / vlan 2 > no ip address > ! > ! > interface Dot11Radio0.2 > encapsulation dot1Q 2 > bridge-group 1 > bridge-group 1 subscriber-loop-control > bridge-group 1 spanning-disabled > bridge-group 1 block-unknown-source > no bridge-group 1 source-learning > no bridge-group 1 unicast-flooding .... if you remove that from "bridge-group 1", and configure ipv4 + ipv6 addresses directly on the dot11radio0.2 interface (making it a routed WiFi network, instead of bridged-to-wired) - will that work? My last test for this failed, because there were no "ipv6 ..." commands at all on "int dot11radio0.2" (while "ip ..." works as expected). Gert Doering -- NetMaster -- Total number of prefixes smaller than registry allocations: 155817 SpaceNet AG Vorstand: Sebastian v. Bomhard Joseph-Dollinger-Bogen 14 Aufsichtsratsvors.: A. Grundner-Culemann D-80807 Muenchen HRB: 136055 (AG Muenchen) Tel: +49 (89) 32356-444 USt-IdNr.: DE813185279 From andy at nosignal.org Wed Aug 4 12:00:34 2010 From: andy at nosignal.org (Andy Davidson) Date: Wed, 4 Aug 2010 11:00:34 +0100 Subject: Cisco 1801W, v6 on wireless, bvi argh. In-Reply-To: <20100804095449.GK73473@Space.Net> References: <24D585B4-F4CD-4CC6-AE98-46C1A3BEE4C7@nosignal.org> <4BF0538D.2000005@netability.ie> <4BF26150.9030407@netability.ie> <8DB354A9-C161-461C-A787-EB7C8EA40E2C@employees.org> <4BF2672D.9080508@netability.ie> <20100804095449.GK73473@Space.Net> Message-ID: <7D6082E3-DD78-4E0F-B2EF-A7CAD90FB6B9@nosignal.org> On 4 Aug 2010, at 10:54, Gert Doering wrote: > and configure ipv4 + ipv6 > addresses directly on the dot11radio0.2 interface (making it a routed > WiFi network, instead of bridged-to-wired) - will that work? > > My last test for this failed, because there were no "ipv6 ..." commands > at all on "int dot11radio0.2" (while "ip ..." works as expected). Hi, Gert The situation does not seem to have changed in this regard. bcliffe-gw(config)#interface Dot11Radio0.2 bcliffe-gw(config-subif)#ipv bcliffe-gw(config-subif)#ipv? % Unrecognized command Andy From lists at quux.de Wed Aug 4 12:00:19 2010 From: lists at quux.de (Jens Link) Date: Wed, 04 Aug 2010 12:00:19 +0200 Subject: Cisco 1801W, v6 on wireless, bvi argh. In-Reply-To: <20100804095449.GK73473@Space.Net> (Gert Doering's message of "Wed, 4 Aug 2010 11:54:49 +0200") References: <24D585B4-F4CD-4CC6-AE98-46C1A3BEE4C7@nosignal.org> <4BF0538D.2000005@netability.ie> <4BF26150.9030407@netability.ie> <8DB354A9-C161-461C-A787-EB7C8EA40E2C@employees.org> <4BF2672D.9080508@netability.ie> <20100804095449.GK73473@Space.Net> Message-ID: <87sk2u3dfw.fsf@oban.berlin.quux.de> Gert Doering writes: > My last test for this failed, because there were no "ipv6 ..." commands > at all on "int dot11radio0.2" (while "ip ..." works as expected). Which IOS version? I had a similar problem on an 871 and upgrading to 12.4(20)T1 helped. cheers Jens -- ------------------------------------------------------------------------- | Foelderichstr. 40 | 13595 Berlin, Germany | +49-151-18721264 | | http://blog.quux.de | jabber: jenslink at guug.de | ------------------- | ------------------------------------------------------------------------- From gert at space.net Wed Aug 4 12:05:38 2010 From: gert at space.net (Gert Doering) Date: Wed, 4 Aug 2010 12:05:38 +0200 Subject: Cisco 1801W, v6 on wireless, bvi argh. In-Reply-To: <87sk2u3dfw.fsf@oban.berlin.quux.de> References: <24D585B4-F4CD-4CC6-AE98-46C1A3BEE4C7@nosignal.org> <4BF0538D.2000005@netability.ie> <4BF26150.9030407@netability.ie> <8DB354A9-C161-461C-A787-EB7C8EA40E2C@employees.org> <4BF2672D.9080508@netability.ie> <20100804095449.GK73473@Space.Net> <87sk2u3dfw.fsf@oban.berlin.quux.de> Message-ID: <20100804100538.GL73473@Space.Net> Hi, On Wed, Aug 04, 2010 at 12:00:19PM +0200, Jens Link wrote: > Gert Doering writes: > > > My last test for this failed, because there were no "ipv6 ..." commands > > at all on "int dot11radio0.2" (while "ip ..." works as expected). > > Which IOS version? > > I had a similar problem on an 871 and upgrading to 12.4(20)T1 helped. "different business unit". (But interesting, I assumed so far that IPv6-on-Wireless wasn't supported on *any* Cisco gear) Gert Doering -- NetMaster -- Total number of prefixes smaller than registry allocations: 155817 SpaceNet AG Vorstand: Sebastian v. Bomhard Joseph-Dollinger-Bogen 14 Aufsichtsratsvors.: A. Grundner-Culemann D-80807 Muenchen HRB: 136055 (AG Muenchen) Tel: +49 (89) 32356-444 USt-IdNr.: DE813185279 From danny at danysek.cz Wed Aug 4 16:00:37 2010 From: danny at danysek.cz (Daniel Suchy) Date: Wed, 04 Aug 2010 16:00:37 +0200 Subject: Cisco 1801W, v6 on wireless, bvi argh. In-Reply-To: <20100804100538.GL73473@Space.Net> References: <24D585B4-F4CD-4CC6-AE98-46C1A3BEE4C7@nosignal.org> <4BF0538D.2000005@netability.ie> <4BF26150.9030407@netability.ie> <8DB354A9-C161-461C-A787-EB7C8EA40E2C@employees.org> <4BF2672D.9080508@netability.ie> <20100804095449.GK73473@Space.Net> <87sk2u3dfw.fsf@oban.berlin.quux.de> <20100804100538.GL73473@Space.Net> Message-ID: <4C597285.80904@danysek.cz> Hello, On 08/04/2010 12:05 PM, Gert Doering wrote: >> I had a similar problem on an 871 and upgrading to 12.4(20)T1 helped. > (But interesting, I assumed so far that IPv6-on-Wireless wasn't supported on > *any* Cisco gear) according to my observations, on C871, with versions up to 12.4(22)T, including maintenance releases, there's no problem with configuring IPv6 on wirelesess interface. With 12.4(24)T, this is problem, probably someone removed (?!) support for IPv6... With regards, Daniel From nick at netability.ie Wed Aug 4 17:22:05 2010 From: nick at netability.ie (Nick Hilliard) Date: Wed, 04 Aug 2010 16:22:05 +0100 Subject: Cisco 1801W, v6 on wireless, bvi argh. In-Reply-To: <4C597285.80904@danysek.cz> References: <24D585B4-F4CD-4CC6-AE98-46C1A3BEE4C7@nosignal.org> <4BF0538D.2000005@netability.ie> <4BF26150.9030407@netability.ie> <8DB354A9-C161-461C-A787-EB7C8EA40E2C@employees.org> <4BF2672D.9080508@netability.ie> <20100804095449.GK73473@Space.Net> <87sk2u3dfw.fsf@oban.berlin.quux.de> <20100804100538.GL73473@Space.Net> <4C597285.80904@danysek.cz> Message-ID: <4C59859D.3050807@netability.ie> On 04/08/2010 15:00, Daniel Suchy wrote: > according to my observations, on C871, with versions up to 12.4(22)T, > including maintenance releases, there's no problem with configuring IPv6 > on wirelesess interface. With 12.4(24)T, this is problem, probably > someone removed (?!) support for IPv6... Which brings up the issue: why was adding ipv6 support to BVIs so difficult in the first place? I would love to know what tortured coding hacks were required to implement this, because sure as hell, if this were a relatively simple matter, Cisco would have implemented it long ago. Nick From gert at space.net Wed Aug 4 17:27:48 2010 From: gert at space.net (Gert Doering) Date: Wed, 4 Aug 2010 17:27:48 +0200 Subject: Cisco 1801W, v6 on wireless, bvi argh. In-Reply-To: <4C59859D.3050807@netability.ie> References: <4BF0538D.2000005@netability.ie> <4BF26150.9030407@netability.ie> <8DB354A9-C161-461C-A787-EB7C8EA40E2C@employees.org> <4BF2672D.9080508@netability.ie> <20100804095449.GK73473@Space.Net> <87sk2u3dfw.fsf@oban.berlin.quux.de> <20100804100538.GL73473@Space.Net> <4C597285.80904@danysek.cz> <4C59859D.3050807@netability.ie> Message-ID: <20100804152748.GV73473@Space.Net> Hi, On Wed, Aug 04, 2010 at 04:22:05PM +0100, Nick Hilliard wrote: > Which brings up the issue: why was adding ipv6 support to BVIs so difficult > in the first place? I would love to know what tortured coding hacks were > required to implement this, because sure as hell, if this were a relatively > simple matter, Cisco would have implemented it long ago. Given that it *still* doesn't work to have routed WiFi interfaces with IPv6, it must be something so horrible that we'll all go back to DECnet or Appletalk if we learn about it... gert -- Total number of prefixes smaller than registry allocations: 155817 SpaceNet AG Vorstand: Sebastian v. Bomhard Joseph-Dollinger-Bogen 14 Aufsichtsratsvors.: A. Grundner-Culemann D-80807 Muenchen HRB: 136055 (AG Muenchen) Tel: +49 (89) 32356-444 USt-IdNr.: DE813185279 From daniel at bit.nl Wed Aug 4 17:40:34 2010 From: daniel at bit.nl (Daniel Verlouw) Date: Wed, 04 Aug 2010 17:40:34 +0200 Subject: Cisco 1801W, v6 on wireless, bvi argh. In-Reply-To: <20100804152748.GV73473@Space.Net> References: <4BF0538D.2000005@netability.ie> <4BF26150.9030407@netability.ie> <8DB354A9-C161-461C-A787-EB7C8EA40E2C@employees.org> <4BF2672D.9080508@netability.ie> <20100804095449.GK73473@Space.Net> <87sk2u3dfw.fsf@oban.berlin.quux.de> <20100804100538.GL73473@Space.Net> <4C597285.80904@danysek.cz> <4C59859D.3050807@netability.ie> <20100804152748.GV73473@Space.Net> Message-ID: <1280936434.1030.26.camel@daniel> On Wed, 2010-08-04 at 17:27 +0200, Gert Doering wrote: > Given that it *still* doesn't work to have routed WiFi interfaces with > IPv6, it must be something so horrible that we'll all go back to DECnet > or Appletalk if we learn about it... well, to make things worse, I've been struggling with a brandnew 887W all day now. Guess what? Routed ipv6 on wifi works perfectly, on SVI wired interfaces it doesn't :( Router advertisements are simply not coming through... "debug ipv6 nd" says everything is fine, RA's are being sent etc. tcpdump on a directly connected host thinks otherwise. *sigh* This is on 15.0.1(M3) btw. --Daniel. From gert at space.net Wed Aug 4 17:42:38 2010 From: gert at space.net (Gert Doering) Date: Wed, 4 Aug 2010 17:42:38 +0200 Subject: Cisco 1801W, v6 on wireless, bvi argh. In-Reply-To: <1280936434.1030.26.camel@daniel> References: <8DB354A9-C161-461C-A787-EB7C8EA40E2C@employees.org> <4BF2672D.9080508@netability.ie> <20100804095449.GK73473@Space.Net> <87sk2u3dfw.fsf@oban.berlin.quux.de> <20100804100538.GL73473@Space.Net> <4C597285.80904@danysek.cz> <4C59859D.3050807@netability.ie> <20100804152748.GV73473@Space.Net> <1280936434.1030.26.camel@daniel> Message-ID: <20100804154238.GX73473@Space.Net> Hi, On Wed, Aug 04, 2010 at 05:40:34PM +0200, Daniel Verlouw wrote: > On Wed, 2010-08-04 at 17:27 +0200, Gert Doering wrote: > > Given that it *still* doesn't work to have routed WiFi interfaces with > > IPv6, it must be something so horrible that we'll all go back to DECnet > > or Appletalk if we learn about it... > > well, to make things worse, I've been struggling with a brandnew 887W > all day now. Guess what? Routed ipv6 on wifi works perfectly, on SVI > wired interfaces it doesn't :( Router advertisements are simply not > coming through... > > "debug ipv6 nd" says everything is fine, RA's are being sent etc. > tcpdump on a directly connected host thinks otherwise. *sigh* > > This is on 15.0.1(M3) btw. Well, *this* is the old "IPv6 on BVI does not work right" bug. Try 15.1T or whatever the original poster reported that worked for his 1800 :-) Gert Doering -- NetMaster -- Total number of prefixes smaller than registry allocations: 155817 SpaceNet AG Vorstand: Sebastian v. Bomhard Joseph-Dollinger-Bogen 14 Aufsichtsratsvors.: A. Grundner-Culemann D-80807 Muenchen HRB: 136055 (AG Muenchen) Tel: +49 (89) 32356-444 USt-IdNr.: DE813185279 From ulrich.kiermayr at univie.ac.at Wed Aug 4 21:17:22 2010 From: ulrich.kiermayr at univie.ac.at (Ulrich Kiermayr) Date: Wed, 4 Aug 2010 21:17:22 +0200 Subject: Cisco 1801W, v6 on wireless, bvi argh. In-Reply-To: References: <24D585B4-F4CD-4CC6-AE98-46C1A3BEE4C7@nosignal.org> <4BF0538D.2000005@netability.ie> <4BF26150.9030407@netability.ie> <8DB354A9-C161-461C-A787-EB7C8EA40E2C@employees.org> <4BF2672D.9080508@netability.ie> Message-ID: <0B0C0EF2-1472-4D2E-BCE2-D3F614EDF3D4@univie.ac.at> On 03.08.2010, at 23:57, Andy Davidson wrote: > > On 18 May 2010, at 11:08, Nick Hilliard wrote: > >> On 18/05/2010 10:58, Ole Troan wrote: >>>>> You didn't try 12.4(22)T, did you? This bug was fixed in that version, and >>>> >>>> I lied - the bug wasn't fixed in that version after all. >>>> >>>> I could have sworn at the time that either this bugid or a similar bugid >>>> was marked fixed in 12.4(22)T. However, looking at the description for >>>> CSCej50923, there's no fixed listed for 12.4T at all. Same with bugid >>>> CSCta27529, which was mentioned in: >>> >>> CSCta27529 is integrated in 15.1T images. >> >> Both CSCej50923 and CSCta27529 are listed as being fixed in 15.1(0.1)T >> onwards. However, if as Andy claims there is no "bridge XXX route ipv6" >> command in 15.1(1)T, then it looks like there's something odd going on. > > Well, I can't explain the following based on the information in the bug tool, but I upgraded to 15.1(2)T (Adv Enterprise), and ipv6 is working on the bvi interface right now. > > bridge XXX router ipv6 still does not appear as a possible command, but I put the ipv6 address on the BVI interface, and everything is singing and dancing as planned. I can confirm tihs for an 877W with 15.1(2)T as well lG uk -- Ulrich Kiermayr Leiter der Abteilung Datennetz & Telefonie Zentraler Informatikdienst der Universitaet Wien Universitaetsstrasse 7, 1010 Wien, AT phone +43 1 4277 14020 jabber xmpp:uk at jabber.univie.ac.at fax +43 1 4277 9140 skype:kiermayr From daniel at bit.nl Thu Aug 5 10:25:32 2010 From: daniel at bit.nl (Daniel Verlouw) Date: Thu, 05 Aug 2010 10:25:32 +0200 Subject: Cisco 1801W, v6 on wireless, bvi argh. In-Reply-To: <20100804154238.GX73473@Space.Net> References: <8DB354A9-C161-461C-A787-EB7C8EA40E2C@employees.org> <4BF2672D.9080508@netability.ie> <20100804095449.GK73473@Space.Net> <87sk2u3dfw.fsf@oban.berlin.quux.de> <20100804100538.GL73473@Space.Net> <4C597285.80904@danysek.cz> <4C59859D.3050807@netability.ie> <20100804152748.GV73473@Space.Net> <1280936434.1030.26.camel@daniel> <20100804154238.GX73473@Space.Net> Message-ID: <1280996732.12246.67.camel@daniel> On Wed, 2010-08-04 at 17:42 +0200, Gert Doering wrote: > > "debug ipv6 nd" says everything is fine, RA's are being sent etc. > > tcpdump on a directly connected host thinks otherwise. *sigh* > > > > This is on 15.0.1(M3) btw. > > Well, *this* is the old "IPv6 on BVI does not work right" bug. really? Mmm, my understanding always has been that this "bug" [1] affected IPv6 on wireless, not on wired as in this case. [1] was it really a bug or just a missing feature? Since 15.1T now specifically mentions IPv6 on BVI: Note that the 887W wireless architecture is much different from let's a 877W. It uses a seperate wifi AP subsystem running its own IOS connecting to the main router subsystem using a GigE switchport. The main router simply sees the AP as "another switchport" and doesn't use BVI's anymore. > Try 15.1T or whatever the original poster reported that worked for his > 1800 :-) actually, it turned out the end-user had a semi-smart-but-braindead switch connected that dropped some multicasts and hence the RA's were not making it all the way through :/ Problem solved, works fine on 15.0 :) --Daniel. From lists at quux.de Thu Aug 5 10:56:57 2010 From: lists at quux.de (Jens Link) Date: Thu, 05 Aug 2010 10:56:57 +0200 Subject: Cisco 1801W, v6 on wireless, bvi argh. In-Reply-To: <87sk2u3dfw.fsf@oban.berlin.quux.de> (Jens Link's message of "Wed, 04 Aug 2010 12:00:19 +0200") References: <24D585B4-F4CD-4CC6-AE98-46C1A3BEE4C7@nosignal.org> <4BF0538D.2000005@netability.ie> <4BF26150.9030407@netability.ie> <8DB354A9-C161-461C-A787-EB7C8EA40E2C@employees.org> <4BF2672D.9080508@netability.ie> <20100804095449.GK73473@Space.Net> <87sk2u3dfw.fsf@oban.berlin.quux.de> Message-ID: <87lj8l5teu.fsf@oban.berlin.quux.de> Jens Link writes: > I had a similar problem on an 871 and upgrading to 12.4(20)T1 helped. Having IPv6 over wireless seems to be a temporary bug in the IOS version above. Looks like they fixed it in 12.4(24)T3 where I can't configure an IPv6 address on the wireless interface. Jens -- ------------------------------------------------------------------------- | Foelderichstr. 40 | 13595 Berlin, Germany | +49-151-18721264 | | http://blog.quux.de | jabber: jenslink at guug.de | ------------------- | ------------------------------------------------------------------------- From gert at space.net Thu Aug 5 14:36:14 2010 From: gert at space.net (Gert Doering) Date: Thu, 5 Aug 2010 14:36:14 +0200 Subject: Cisco 1801W, v6 on wireless, bvi argh. In-Reply-To: <1280996732.12246.67.camel@daniel> References: <20100804095449.GK73473@Space.Net> <87sk2u3dfw.fsf@oban.berlin.quux.de> <20100804100538.GL73473@Space.Net> <4C597285.80904@danysek.cz> <4C59859D.3050807@netability.ie> <20100804152748.GV73473@Space.Net> <1280936434.1030.26.camel@daniel> <20100804154238.GX73473@Space.Net> <1280996732.12246.67.camel@daniel> Message-ID: <20100805123614.GC73473@Space.Net> Hi, On Thu, Aug 05, 2010 at 10:25:32AM +0200, Daniel Verlouw wrote: > On Wed, 2010-08-04 at 17:42 +0200, Gert Doering wrote: > > > "debug ipv6 nd" says everything is fine, RA's are being sent etc. > > > tcpdump on a directly connected host thinks otherwise. *sigh* > > > > > > This is on 15.0.1(M3) btw. > > > > Well, *this* is the old "IPv6 on BVI does not work right" bug. > > really? Mmm, my understanding always has been that this "bug" [1] > affected IPv6 on wireless, not on wired as in this case. In my tests, it affected both. You could configure IPv6 on the BVI, but the box would get all confused about it, and would not recognize the network as "directly connected" and send packets to it. Bridging IPv6 packets from *other* nodes ("Node A on WiFi, node B on wirde LAN") worked fine, just talking to the Cisco's BVI didn't work. > [1] was it really a bug or just a missing feature? Since 15.1T now > specifically mentions IPv6 on BVI: > Well, since you could configure it, and it didn't work, it's a bug in my book. Of course one can always relabel bugs as "missing features". > Note that the 887W wireless architecture is much different from let's a > 877W. It uses a seperate wifi AP subsystem running its own IOS > connecting to the main router subsystem using a GigE switchport. The > main router simply sees the AP as "another switchport" and doesn't use > BVI's anymore. Ah. I didn't know that. (I long lost track on the plethora of 8xx ciscos that are all weird in their own way, like the 861 that is said to never ever get IPv6 support in any way "because it is not powerful enough"). > > Try 15.1T or whatever the original poster reported that worked for his > > 1800 :-) > > actually, it turned out the end-user had a semi-smart-but-braindead > switch connected that dropped some multicasts and hence the RA's were > not making it all the way through :/ Problem solved, works fine on > 15.0 :) Glad that it works now - and annyoing to learn that there is now some more L2 gear that breaks IPv6... Gert Doering -- NetMaster -- Total number of prefixes smaller than registry allocations: 155817 SpaceNet AG Vorstand: Sebastian v. Bomhard Joseph-Dollinger-Bogen 14 Aufsichtsratsvors.: A. Grundner-Culemann D-80807 Muenchen HRB: 136055 (AG Muenchen) Tel: +49 (89) 32356-444 USt-IdNr.: DE813185279 -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 306 bytes Desc: not available Url : http://lists.cluenet.de/pipermail/ipv6-ops/attachments/20100805/49833d37/attachment.bin From me at benedikt-stockebrand.de Thu Aug 5 15:56:17 2010 From: me at benedikt-stockebrand.de (Benedikt Stockebrand) Date: Thu, 05 Aug 2010 13:56:17 +0000 Subject: Cisco 1801W, v6 on wireless, bvi argh. In-Reply-To: <20100805123614.GC73473@Space.Net> (Gert Doering's message of "Thu, 5 Aug 2010 14:36:14 +0200") References: <20100804095449.GK73473@Space.Net> <87sk2u3dfw.fsf@oban.berlin.quux.de> <20100804100538.GL73473@Space.Net> <4C597285.80904@danysek.cz> <4C59859D.3050807@netability.ie> <20100804152748.GV73473@Space.Net> <1280936434.1030.26.camel@daniel> <20100804154238.GX73473@Space.Net> <1280996732.12246.67.camel@daniel> <20100805123614.GC73473@Space.Net> Message-ID: Hi Gert and list, Gert Doering writes: > Ah. I didn't know that. (I long lost track on the plethora of 8xx > ciscos that are all weird in their own way, like the 861 that is said > to never ever get IPv6 support in any way "because it is not powerful > enough"). if "not powerful enough" actually means "not enough RAM and/or flash memory" and they are talking about dual stacking rather than *replacing* IPv4 with IPv6, their statement actually makes sense. The online data sheet hints that the 860 series has both flash and RAM soldered on-board, so they are designed for a mostly fixed set of functionalities. If that didn't include IPv6 from the beginning, then retrofitting it can be infeasible. If you plan to gain weight, don't buy tight-fitting pants. Cheers, Benedikt -- Business Grade IPv6 Consulting, Training, Projects Benedikt Stockebrand, Dipl.-Inform. http://www.benedikt-stockebrand.de/ From gert at space.net Thu Aug 5 16:03:09 2010 From: gert at space.net (Gert Doering) Date: Thu, 5 Aug 2010 16:03:09 +0200 Subject: Cisco 1801W, v6 on wireless, bvi argh. In-Reply-To: References: <87sk2u3dfw.fsf@oban.berlin.quux.de> <20100804100538.GL73473@Space.Net> <4C597285.80904@danysek.cz> <4C59859D.3050807@netability.ie> <20100804152748.GV73473@Space.Net> <1280936434.1030.26.camel@daniel> <20100804154238.GX73473@Space.Net> <1280996732.12246.67.camel@daniel> <20100805123614.GC73473@Space.Net> Message-ID: <20100805140309.GI73473@Space.Net> Hi, On Thu, Aug 05, 2010 at 01:56:17PM +0000, Benedikt Stockebrand wrote: > Gert Doering writes: > > > Ah. I didn't know that. (I long lost track on the plethora of 8xx > > ciscos that are all weird in their own way, like the 861 that is said > > to never ever get IPv6 support in any way "because it is not powerful > > enough"). > > if "not powerful enough" actually means "not enough RAM and/or flash > memory" and they are talking about dual stacking rather than > *replacing* IPv4 with IPv6, their statement actually makes sense. Well. It's not like the 861 is an old product - there's 800 series routers that are much older and do IPv6 just fine, like the 836... So this is more like "how can a vendor position themselves as 'premium vendor' and 'we do IPv6 everywhere!' and at the same time design a product in, what, 2005?, that cannot do IPv6?" It's not like this is a Linksys US$ 30 box... Gert Doering -- NetMaster -- Total number of prefixes smaller than registry allocations: 155817 SpaceNet AG Vorstand: Sebastian v. Bomhard Joseph-Dollinger-Bogen 14 Aufsichtsratsvors.: A. Grundner-Culemann D-80807 Muenchen HRB: 136055 (AG Muenchen) Tel: +49 (89) 32356-444 USt-IdNr.: DE813185279 From ulrich.kiermayr at univie.ac.at Thu Aug 5 16:06:57 2010 From: ulrich.kiermayr at univie.ac.at (Ulrich Kiermayr) Date: Thu, 5 Aug 2010 16:06:57 +0200 Subject: Cisco 1801W, v6 on wireless, bvi argh. In-Reply-To: References: <20100804095449.GK73473@Space.Net> <87sk2u3dfw.fsf@oban.berlin.quux.de> <20100804100538.GL73473@Space.Net> <4C597285.80904@danysek.cz> <4C59859D.3050807@netability.ie> <20100804152748.GV73473@Space.Net> <1280936434.1030.26.camel@daniel> <20100804154238.GX73473@Space.Net> <1280996732.12246.67.camel@daniel> <20100805123614.GC73473@Space.Net> Message-ID: <2414A3EC-C4F9-47DA-9EE5-381787F5A9DD@univie.ac.at> On 05.08.2010, at 15:56, Benedikt Stockebrand wrote: > Hi Gert and list, > > Gert Doering writes: > >> Ah. I didn't know that. (I long lost track on the plethora of 8xx >> ciscos that are all weird in their own way, like the 861 that is said >> to never ever get IPv6 support in any way "because it is not powerful >> enough"). > > if "not powerful enough" actually means "not enough RAM and/or flash > memory" and they are talking about dual stacking rather than > *replacing* IPv4 with IPv6, their statement actually makes sense. > > The online data sheet hints that the 860 series has both flash and RAM > soldered on-board, so they are designed for a mostly fixed set of > functionalities. If that didn't include IPv6 from the beginning, then > retrofitting it can be infeasible. The Flash is not so much of an issue - as long as you have an TFTP at hand ;-). At least i had to do it that way for my 877, which also comes with to little flash for all those nice features. lG uk -- Ulrich Kiermayr jabber xmpp:uk at jabber.univie.ac.at Leiter der Abteilung Datennetz und Telefonie skype:kiermayr Vienna University Computer Center phone +43 1 4277 14020 Universitaetsstrasse 7, 1010 Wien, AT fax +43 1 4277 9140 -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 3141 bytes Desc: not available Url : http://lists.cluenet.de/pipermail/ipv6-ops/attachments/20100805/1ba133da/attachment.bin From frnkblk at iname.com Thu Aug 5 16:19:45 2010 From: frnkblk at iname.com (Frank Bulk - iName.com) Date: Thu, 5 Aug 2010 09:19:45 -0500 Subject: Cisco 1801W, v6 on wireless, bvi argh. In-Reply-To: <20100805140309.GI73473@Space.Net> References: <87sk2u3dfw.fsf@oban.berlin.quux.de> <20100804100538.GL73473@Space.Net> <4C597285.80904@danysek.cz> <4C59859D.3050807@netability.ie> <20100804152748.GV73473@Space.Net> <1280936434.1030.26.camel@daniel> <20100804154238.GX73473@Space.Net> <1280996732.12246.67.camel@daniel> <20100805123614.GC73473@Space.Net> <20100805140309.GI73473@Space.Net> Message-ID: I opened a case a few months ago about missing support for handing IPv6 DNS server IP addresses for SSL VPN connection connections on the ASA. The support tech opened up a feature request, and I just followed up yesterday, including this URL which is a video clip from John Chambers about IPv6: http://www.youtube.com/watch?v=VCCgVh8wFdA Who knows if it will help. =) Frank -----Original Message----- From: ipv6-ops-bounces+frnkblk=iname.com at lists.cluenet.de [mailto:ipv6-ops-bounces+frnkblk=iname.com at lists.cluenet.de] On Behalf Of Gert Doering Sent: Thursday, August 05, 2010 9:03 AM To: Benedikt Stockebrand Cc: ipv6-ops at lists.cluenet.de Subject: Re: Cisco 1801W, v6 on wireless, bvi argh. So this is more like "how can a vendor position themselves as 'premium vendor' and 'we do IPv6 everywhere!' and at the same time design a product in, what, 2005?, that cannot do IPv6?" It's not like this is a Linksys US$ 30 box... Gert Doering -- NetMaster From frnkblk at iname.com Sun Aug 8 01:22:19 2010 From: frnkblk at iname.com (Frank Bulk - iName.com) Date: Sat, 7 Aug 2010 18:22:19 -0500 Subject: Cisco configuration question Message-ID: I was able to get a configuration with SLAAC and DHCP-PD going, like this: ----------- | 7206VXR | local DHCP pool ----------- | SLAAC using single /64 | gateway/router (LAN address space obtained via DHCP-PD, | SLAAC DNS suffix and server via stateless DHCP) Laptop But I'd like to try the following: ----------- | 7206VXR | local DHCP pools for DHCP-PD and gateway/ ----------- router WAN | DHCP pool, each gateway/router in separate /64 | gateway/router (LAN address space obtained via DHCP-PD, | SLAAC DNS suffix and server via stateless DHCP) Laptop Is that even possible? Frank From nanog at 85d5b20a518b8f6864949bd940457dc124746ddc.nosense.org Sun Aug 8 03:24:13 2010 From: nanog at 85d5b20a518b8f6864949bd940457dc124746ddc.nosense.org (Mark Smith) Date: Sun, 8 Aug 2010 10:54:13 +0930 Subject: Cisco configuration question In-Reply-To: References: Message-ID: <20100808105413.77c07539@opy.nosense.org> On Sat, 7 Aug 2010 18:22:19 -0500 "Frank Bulk - iName.com" wrote: > I was able to get a configuration with SLAAC and DHCP-PD going, like this: > > ----------- > | 7206VXR | local DHCP pool > ----------- > | SLAAC using single /64 > | > gateway/router (LAN address space obtained via DHCP-PD, > | SLAAC DNS suffix and server via stateless DHCP) > Laptop > > But I'd like to try the following: > > ----------- > | 7206VXR | local DHCP pools for DHCP-PD and gateway/ > ----------- router WAN > | DHCP pool, each gateway/router in separate /64 > | > gateway/router (LAN address space obtained via DHCP-PD, > | SLAAC DNS suffix and server via stateless DHCP) > Laptop > > Is that even possible? > Yes, with the caveat that I've seen it work on an ASR. You have two local IPv6 pools on the 7206VXR. One is used for /64s for the links to the gateway/router, the other is used for DHCPv6-PD assignments. To specify the pool to use for the /64 assignments for the links to the gateways/routers, you use the Cisco RADIUS VSA Framed-IPv6-Pool attribute - http://www.cisco.com/en/US/docs/ios/ios_xe/ipv6/configuration/guide/ip6-adsl_dial_xe.html#wp1075244 I'm not sure if there is a local CLI equivalent identifying the default pool to use, that you'd specify under the virtual template. Regards, Mark. From me at benedikt-stockebrand.de Sun Aug 8 12:40:56 2010 From: me at benedikt-stockebrand.de (Benedikt Stockebrand) Date: Sun, 08 Aug 2010 10:40:56 +0000 Subject: Cisco 1801W, v6 on wireless, bvi argh. In-Reply-To: <20100805140309.GI73473@Space.Net> (Gert Doering's message of "Thu, 5 Aug 2010 16:03:09 +0200") References: <87sk2u3dfw.fsf@oban.berlin.quux.de> <20100804100538.GL73473@Space.Net> <4C597285.80904@danysek.cz> <4C59859D.3050807@netability.ie> <20100804152748.GV73473@Space.Net> <1280936434.1030.26.camel@daniel> <20100804154238.GX73473@Space.Net> <1280996732.12246.67.camel@daniel> <20100805123614.GC73473@Space.Net> <20100805140309.GI73473@Space.Net> Message-ID: Hi Gert and list, > Well. It's not like the 861 is an old product - there's 800 series > routers that are much older and do IPv6 just fine, like the 836... yes, but the early models were quite likely designed with a less specific usage scenario in mind. When a vendor first starts to build products for a market they are unfamiliar with (in this case SoHo class routers), they try to keep them flexible so they can serve a large market. Once they understand the market better and have established some reputation, they start to "optimize" their products, possibly forking multiple product lines to serve different segments of the market. For that reason, whenever you buy a successor model to replace some gear it always pays to check that all the features you need are still available with the new one---especially when the new model is cheaper than the old one. Classic examples are the Adaptec 2940U/2940AU debacle years ago, the SUNs move from E250 (server board) to E220R (workstation board) and of course the assorted Linksys WRT54* products. > So this is more like "how can a vendor position themselves as 'premium > vendor' and 'we do IPv6 everywhere!' and at the same time design a > product in, what, 2005?, that cannot do IPv6?" Did Cisco actually claim to "do IPv6 everywhere" as early as 2005? As far as I remember their IPv6 support at that time it was rather limited, and considering the discussion about the 18xx series WLAN interfaces they still aren't really up to speed today. > It's not like this is a Linksys US$ 30 box... No, the extra money you pay is for (some kind of) support, possibly/hopefully better quality components and, most importantly, proper lifetime management with an extended availability guarantee for additional machines and spare parts, early on end-of-life announcements and such. Try to order an exact replacement for a two year^W^Wsix month old Linksys router; chances are that you are told to try to buy one second hand on ebay if you don't want one of the current models. But otherwise it *is* like a Linksys box, only at a slightly higher level. Unless a feature is relevant to a significant part of the intended market, you don't implement it to keep the price low. And in 2005 very few people cared about spending money on IPv6, so spending money to support IPv6 or even just to add the memory to allow IPv6 to be retrofitted later on would have been counterproductive from a business point of view. Cheers, Benedikt -- Business Grade IPv6 Consulting, Training, Projects Benedikt Stockebrand, Dipl.-Inform. http://www.benedikt-stockebrand.de/ From tore.anderson at redpill-linpro.com Thu Aug 12 16:33:12 2010 From: tore.anderson at redpill-linpro.com (Tore Anderson) Date: Thu, 12 Aug 2010 16:33:12 +0200 Subject: IPv6-capable 10G switches Message-ID: <4C640628.1000502@redpill-linpro.com> Hi guys, I need a couple of new core switches for a medium-sized data centre I'm setting up soon, and I'm curious to hear if any of you have any recommendations or tips on suitable products? The budget is limited, maybe around US$15.000 per unit max, so I'm focusing on the fixed-configuration models for now. My requirements are at least 20 SFP+ ports and full L3 support for both IPv6 and IPv4 (fast-path IPv6, OSPFv3, VRRPv3, filters). On the nice-to-have list are things like candidate configuration database with commit/rollback capabilities, in-service software upgrades, MPLS, and stacking support. Disappointingly enough I've only found two switches that appear to satisfy my requirements so far: the Extreme Summit X650 and the BNT RackSwitch G8124. The latter is completely unknown to me, but I do have some Extremes already and I can't say I'm a very big XOS fan... The other models I've looked at and rejected are: Arista 7124S (no IPv6), Brocade TurboIron 24X (no IPv6), Cisco Catalyst 4900M (X2 optics), Cisco Nexus 5010 (no layer 3), Force10 S2410 (no IPv6), and Juniper EX 4500 (no IPv6). Best regards, -- Tore Anderson Redpill Linpro AS - http://www.redpill-linpro.com/ Tel: +47 21 54 41 27 From nick at foobar.org Thu Aug 12 16:50:11 2010 From: nick at foobar.org (Nick Hilliard) Date: Thu, 12 Aug 2010 15:50:11 +0100 Subject: IPv6-capable 10G switches In-Reply-To: <4C640628.1000502@redpill-linpro.com> References: <4C640628.1000502@redpill-linpro.com> Message-ID: <4C640A23.6080806@foobar.org> On 12/08/2010 15:33, Tore Anderson wrote: > I need a couple of new core switches for a medium-sized data centre I'm > setting up soon, and I'm curious to hear if any of you have any > recommendations or tips on suitable products? The budget is limited, > maybe around US$15.000 per unit max, so I'm focusing on the > fixed-configuration models for now. > > My requirements are at least 20 SFP+ ports and full L3 support for both > IPv6 and IPv4 (fast-path IPv6, OSPFv3, VRRPv3, filters). On the > nice-to-have list are things like candidate configuration database with > commit/rollback capabilities, in-service software upgrades, MPLS, and > stacking support. > > Disappointingly enough I've only found two switches that appear to > satisfy my requirements so far: the Extreme Summit X650 and the BNT > RackSwitch G8124. The latter is completely unknown to me, but I do have > some Extremes already and I can't say I'm a very big XOS fan... Brocade have ipv6 support on the roadmap for the TI24X. However, it won't do full routes; istr that it only has enough fib capacity for a couple of thousand prefixes. It also won't do BGP or MPLS. The X650 supports 12k prefixes. You may want to revisit your requirements. There are currently no L3 switches out there which fit your bill at that price range. Maybe if you tripled or quadrupled your budget, you'd get something, but not at $15k. The Juniper EX2500 is a rebadged G8124. If you have access to J documentation, you should be able to read up on its specs. Nick From kuenzler at init7.net Thu Aug 12 17:04:40 2010 From: kuenzler at init7.net (Fredy Kuenzler) Date: Thu, 12 Aug 2010 17:04:40 +0200 Subject: IPv6-capable 10G switches In-Reply-To: <4C640628.1000502@redpill-linpro.com> References: <4C640628.1000502@redpill-linpro.com> Message-ID: <4C640D88.1090101@init7.net> Am 12.08.2010 16:33, schrieb Tore Anderson: > I need a couple of new core switches for a medium-sized data centre I'm > setting up soon, and I'm curious to hear if any of you have any > recommendations or tips on suitable products? The budget is limited, > maybe around US$15.000 per unit max, so I'm focusing on the > fixed-configuration models for now. > > My requirements are at least 20 SFP+ ports and full L3 support for both > IPv6 and IPv4 (fast-path IPv6, OSPFv3, VRRPv3, filters). On the > nice-to-have list are things like candidate configuration database with > commit/rollback capabilities, in-service software upgrades, MPLS, and > stacking support. > > Disappointingly enough I've only found two switches that appear to > satisfy my requirements so far: the Extreme Summit X650 and the BNT > RackSwitch G8124. The latter is completely unknown to me, but I do have > some Extremes already and I can't say I'm a very big XOS fan... > > The other models I've looked at and rejected are: Arista 7124S (no > IPv6), Brocade TurboIron 24X (no IPv6), Cisco Catalyst 4900M (X2 optics), > Cisco Nexus 5010 (no layer 3), Force10 S2410 (no IPv6), and Juniper EX > 4500 (no IPv6). Check out the Brocade CER 2000 series. We have a number of these boxes operational and are happy with them. - 1 unit box - 24 or 48* GigE, copper or fibre, 2*10gig optional - full BGP tables, OSPFv3, v4 and v6 (v6 from 5.x and higher firmware) - MPLS - look & feel like the large Brocade boxes such as XMR and MLX not sure about the VRRPv3, though... You should be able to get one without 10gig ports for less than $15000 (street price). Hope this helps. Fredy K?nzler Init7 / AS13030 From tore.anderson at redpill-linpro.com Thu Aug 12 17:25:41 2010 From: tore.anderson at redpill-linpro.com (Tore Anderson) Date: Thu, 12 Aug 2010 17:25:41 +0200 Subject: IPv6-capable 10G switches In-Reply-To: <4C640A23.6080806@foobar.org> References: <4C640628.1000502@redpill-linpro.com> <4C640A23.6080806@foobar.org> Message-ID: <4C641275.3010704@redpill-linpro.com> Hi Nick, * Nick Hilliard > Brocade have ipv6 support on the roadmap for the TI24X. However, it won't > do full routes; istr that it only has enough fib capacity for a couple of > thousand prefixes. It also won't do BGP or MPLS. > > The X650 supports 12k prefixes. I connect to transits and peers elsewhere so routing table constraints and lack of BGP support isn't a problem. Do you happen to know when the roadmap says IPv6 is scheduled to be ready on the TI24X? > You may want to revisit your requirements. There are currently no L3 > switches out there which fit your bill at that price range. Maybe if you > tripled or quadrupled your budget, you'd get something, but not at $15k. Well, the G8124 lists at less than $15k (and I don't know anyone who pays full list price for anything:) and it does appear to meet my requirements. Don't know about the nice-to-have items though - it would be nice to hear from somebody that has first-hand experience with running it in production. You'll get the Arista for less than $15k too, and IPv6 is supposedly planned for a future software release. No idea of when though, and you never know if it will require a feature licence once it arrives... > The Juniper EX2500 is a rebadged G8124. If you have access to J > documentation, you should be able to read up on its specs. I know, but as far as I know Juniper has stripped away all L3 features from it so it's not really an option. Thanks for your reply, -- Tore Anderson Redpill Linpro AS - http://www.redpill-linpro.com/ Tel: +47 21 54 41 27 From ghankins at mindspring.com Thu Aug 12 17:31:10 2010 From: ghankins at mindspring.com (Greg Hankins) Date: Thu, 12 Aug 2010 11:31:10 -0400 Subject: IPv6-capable 10G switches In-Reply-To: <4C640D88.1090101@init7.net> References: <4C640628.1000502@redpill-linpro.com> <4C640D88.1090101@init7.net> Message-ID: <20100812153110.GA1427@brocade.com> On Thu, Aug 12, 2010 at 05:04:40PM +0200, Fredy Kuenzler wrote: >Check out the Brocade CER 2000 series. We have a number of these boxes >operational and are happy with them. > >- 1 unit box >- 24 or 48* GigE, copper or fibre, 2*10gig optional >- full BGP tables, OSPFv3, v4 and v6 (v6 from 5.x and higher firmware) >- MPLS >- look & feel like the large Brocade boxes such as XMR and MLX >not sure about the VRRPv3, though... VRRPv3 will be available for the NetIron product line (CES, CER, MLX, XMR) in release 5.1 next month. Greg (disclaimer: I work for Brocade) -- Greg Hankins From martin at millnert.se Thu Aug 12 18:29:28 2010 From: martin at millnert.se (Martin Millnert) Date: Thu, 12 Aug 2010 18:29:28 +0200 Subject: IPv6-capable 10G switches In-Reply-To: <4C640D88.1090101@init7.net> References: <4C640628.1000502@redpill-linpro.com> <4C640D88.1090101@init7.net> Message-ID: <1281630568.13897.471.camel@hsa.vpn.anti> On Thu, 2010-08-12 at 17:04 +0200, Fredy Kuenzler wrote: > - 24 or 48* GigE, copper or fibre, 2*10gig optional That unfortunately fails the "at least 20 SFP+" requirement of Tore. Regards, -- Martin Millnert -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 198 bytes Desc: This is a digitally signed message part Url : http://lists.cluenet.de/pipermail/ipv6-ops/attachments/20100812/4967a4f5/attachment.bin From marcoh at marcoh.net Thu Aug 12 19:13:26 2010 From: marcoh at marcoh.net (Marco Hogewoning) Date: Thu, 12 Aug 2010 19:13:26 +0200 Subject: IPv6-capable 10G switches In-Reply-To: <4C641275.3010704@redpill-linpro.com> References: <4C640628.1000502@redpill-linpro.com> <4C640A23.6080806@foobar.org> <4C641275.3010704@redpill-linpro.com> Message-ID: > You'll get the Arista for less than $15k too, and IPv6 is supposedly > planned for a future software release. No idea of when though, and you > never know if it will require a feature licence once it arrives... Atrista will take another year or so to come up with IPv6 MarcoH From kuenzler at init7.net Thu Aug 12 22:04:27 2010 From: kuenzler at init7.net (Fredy Kuenzler) Date: Thu, 12 Aug 2010 22:04:27 +0200 Subject: IPv6-capable 10G switches In-Reply-To: <1281630568.13897.471.camel@hsa.vpn.anti> References: <4C640628.1000502@redpill-linpro.com> <4C640D88.1090101@init7.net> <1281630568.13897.471.camel@hsa.vpn.anti> Message-ID: <4C6453CB.5080801@init7.net> Am 12.08.2010 18:29, schrieb Martin Millnert: > On Thu, 2010-08-12 at 17:04 +0200, Fredy Kuenzler wrote: >> - 24 or 48* GigE, copper or fibre, 2*10gig optional > > That unfortunately fails the "at least 20 SFP+" requirement of Tore. Don't think so, depends on the model. Either 20 SFP + 4 combo ports, or 20 copper + 4 combo. And the same with total 48 ports. F. From swmike at swm.pp.se Thu Aug 12 22:11:12 2010 From: swmike at swm.pp.se (Mikael Abrahamsson) Date: Thu, 12 Aug 2010 22:11:12 +0200 (CEST) Subject: IPv6-capable 10G switches In-Reply-To: <4C6453CB.5080801@init7.net> References: <4C640628.1000502@redpill-linpro.com> <4C640D88.1090101@init7.net> <1281630568.13897.471.camel@hsa.vpn.anti> <4C6453CB.5080801@init7.net> Message-ID: On Thu, 12 Aug 2010, Fredy Kuenzler wrote: > Don't think so, depends on the model. Either 20 SFP + 4 combo ports, or > 20 copper + 4 combo. And the same with total 48 ports. SFP+ is 10GE, not GE. -- Mikael Abrahamsson email: swmike at swm.pp.se From kuenzler at init7.net Fri Aug 13 07:57:45 2010 From: kuenzler at init7.net (Fredy Kuenzler) Date: Fri, 13 Aug 2010 07:57:45 +0200 Subject: IPv6-capable 10G switches In-Reply-To: References: <4C640628.1000502@redpill-linpro.com> <4C640D88.1090101@init7.net> <1281630568.13897.471.camel@hsa.vpn.anti> <4C6453CB.5080801@init7.net> Message-ID: <4C64DED9.7090308@init7.net> Am 12.08.2010 22:11, schrieb Mikael Abrahamsson: >> Don't think so, depends on the model. Either 20 SFP + 4 combo ports, or >> 20 copper + 4 combo. And the same with total 48 ports. > > SFP+ is 10GE, not GE. Oh, I must have overlooked the +. My apologies. But 20*10gig for $15000 including Layer-3 support etc. seems to be very unreasonable to me... I'd like to know which vendor is selling such gear for this price. F. From Sam.Wilson at ed.ac.uk Fri Aug 13 11:29:45 2010 From: Sam.Wilson at ed.ac.uk (Sam Wilson) Date: Fri, 13 Aug 2010 10:29:45 +0100 Subject: IPv6-capable 10G switches In-Reply-To: <4C640628.1000502@redpill-linpro.com> References: <4C640628.1000502@redpill-linpro.com> Message-ID: On 12 Aug 2010, at 15:33, Tore Anderson wrote: > Hi guys, > > I need a couple of new core switches for a medium-sized data centre > I'm > setting up soon, and I'm curious to hear if any of you have any > recommendations or tips on suitable products? The budget is limited, > maybe around US$15.000 per unit max, so I'm focusing on the > fixed-configuration models for now. > > My requirements are at least 20 SFP+ ports and full L3 support for > both > IPv6 and IPv4 (fast-path IPv6, OSPFv3, VRRPv3, filters). On the > nice-to-have list are things like candidate configuration database > with > commit/rollback capabilities, in-service software upgrades, MPLS, and > stacking support. > > Disappointingly enough I've only found two switches that appear to > satisfy my requirements so far: the Extreme Summit X650 and the BNT > RackSwitch G8124. The latter is completely unknown to me, but I do > have > some Extremes already and I can't say I'm a very big XOS fan... We have a couple of G8124s but they are used entirely as rack-top L2 switches - I have no experience of L3 configuration or performance with them. The CLI is Cisco-like but not very - the logic of how you configure features is very different. (If you have experience of how "like", say, 3Com 5500s are to Ciscos you'll know how "like" the BNTs are, but in yet another completely different way.) That said, even though we're not stretching their capabilities the G8124s seem to do what you tell them to; we've not found any obvious glitches or bugs. We are using Cisco compatible PVST+ to run different spanning tree topologies and Blade seem to have got that right. The folks with 10G connected servers are having some problems, but that seems to be a server issue rather than a switch problem. Not very useful, I'm afraid, but perhaps it might help a little. Sam Sam Wilson Network Team, IT Infrastructure Information Services, The University of Edinburgh Edinburgh, Scotland, UK -- The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336. From tore.anderson at redpill-linpro.com Tue Aug 17 15:40:49 2010 From: tore.anderson at redpill-linpro.com (Tore Anderson) Date: Tue, 17 Aug 2010 15:40:49 +0200 Subject: IPv6-capable 10G switches In-Reply-To: <4C640628.1000502@redpill-linpro.com> References: <4C640628.1000502@redpill-linpro.com> Message-ID: <4C6A9161.50105@redpill-linpro.com> I got a few suggestions off list I thought I'd share with you for completeness: 1) H3C S5820X-28S - although I'm rather sceptical about the current software quality of the Chinese manufacturers, judging from the data sheet it does appear to have quite complete IPv6 support. 2) Dell PowerConnect 8024F. In my experience Dell cares very little about software maintenance, so I'm a bit sceptical about using this one in a core role too, but like the H3C it does appear to have decent IPv6 support. Anyone know who is the OEM for this box, by the way? Entrasys was mentioned too, but I failed to find a model on their web page that satisfies my requirements. I also spoke with an Arista rep, who could confirm what Marco was saying - IPv6 support is currently scheduled for 2H2011. But he did also hint that it might be possible to get it bumped to 1H2011 if necessary to win some business. Best regards, -- Tore Anderson Redpill Linpro AS - http://www.redpill-linpro.com/ Tel: +47 21 54 41 27 From ulrich.kiermayr at univie.ac.at Thu Aug 19 11:22:55 2010 From: ulrich.kiermayr at univie.ac.at (Ulrich Kiermayr) Date: Thu, 19 Aug 2010 11:22:55 +0200 Subject: Google over IPv6 Message-ID: <4183E1EF-97E2-451F-82B6-D050CE1EFD09@univie.ac.at> Hello List, Since we were trying to reach Google regarding IPv6 (http://www.google.com/intl/en/ipv6/) for a while now without any success. Does anyone have an alternate way of reaching them, except what is mentioned on their webpage? lG uk -- Ulrich Kiermayr jabber xmpp:uk at jabber.univie.ac.at Leiter der Abteilung Datennetz und Telefonie skype:kiermayr Vienna University Computer Center phone +43 1 4277 14020 Universitaetsstrasse 7, 1010 Wien, AT fax +43 1 4277 9140 From sesse at google.com Thu Aug 19 11:29:01 2010 From: sesse at google.com (Steinar H. Gunderson) Date: Thu, 19 Aug 2010 11:29:01 +0200 Subject: Google over IPv6 In-Reply-To: <4183E1EF-97E2-451F-82B6-D050CE1EFD09@univie.ac.at> References: <4183E1EF-97E2-451F-82B6-D050CE1EFD09@univie.ac.at> Message-ID: Den 19. august 2010 11:22 skrev Ulrich Kiermayr f?lgende: > Since we were trying to reach Google regarding IPv6 (http://www.google.com/intl/en/ipv6/) for a while now without any success. Does anyone have an alternate way of reaching them, except what is mentioned on their webpage? Hi Ulrich, We're sorry that people sometimes fall through the cracks; there are a lot of requests, and sometimes, as in your case, they get stalled all too long. We've changed the process a bit, though ? first, note the new policy at http://www.google.com/intl/en/ipv6/faq.html#request . Second, given that you meet the updated requirements, you might want to give the new form at https://services.google.com/fb/forms/requestipv6/ a try, as it captures all the information we need to evaluate your request. /* Steinar */ -- Software Engineer, Google Switzerland From mbertozzi at yahoo.com Thu Aug 19 12:10:22 2010 From: mbertozzi at yahoo.com (Matteo Bertozzi) Date: Thu, 19 Aug 2010 03:10:22 -0700 (PDT) Subject: Google over IPv6 In-Reply-To: <4183E1EF-97E2-451F-82B6-D050CE1EFD09@univie.ac.at> References: <4183E1EF-97E2-451F-82B6-D050CE1EFD09@univie.ac.at> Message-ID: <181013.16512.qm@web55403.mail.re4.yahoo.com> Hello Ulrich, I've tried http://ipv6.google.com and it works very well, even part of links inside the page are still on v4 site. Rgds Matteo ________________________________ From: Ulrich Kiermayr To: ipv6-ops at lists.cluenet.de Sent: Thu, August 19, 2010 11:22:55 AM Subject: Google over IPv6 Hello List, Since we were trying to reach Google regarding IPv6 (http://www.google.com/intl/en/ipv6/) for a while now without any success. Does anyone have an alternate way of reaching them, except what is mentioned on their webpage? lG uk -- Ulrich Kiermayr jabber xmpp:uk at jabber.univie.ac.at Leiter der Abteilung Datennetz und Telefonie skype:kiermayr Vienna University Computer Center phone +43 1 4277 14020 Universitaetsstrasse 7, 1010 Wien, AT fax +43 1 4277 9140 -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.cluenet.de/pipermail/ipv6-ops/attachments/20100819/e83003c4/attachment.html From gert at space.net Thu Aug 19 12:25:34 2010 From: gert at space.net (Gert Doering) Date: Thu, 19 Aug 2010 12:25:34 +0200 Subject: Google over IPv6 In-Reply-To: <181013.16512.qm@web55403.mail.re4.yahoo.com> References: <4183E1EF-97E2-451F-82B6-D050CE1EFD09@univie.ac.at> <181013.16512.qm@web55403.mail.re4.yahoo.com> Message-ID: <20100819102534.GK61734@Space.Net> Hi, On Thu, Aug 19, 2010 at 03:10:22AM -0700, Matteo Bertozzi wrote: > I've tried http://ipv6.google.com and it works very well, even part of links > inside the page are still on v4 site. That's not the point. Gert Doering -- NetMaster -- Total number of prefixes smaller than registry allocations: 155817 SpaceNet AG Vorstand: Sebastian v. Bomhard Joseph-Dollinger-Bogen 14 Aufsichtsratsvors.: A. Grundner-Culemann D-80807 Muenchen HRB: 136055 (AG Muenchen) Tel: +49 (89) 32356-444 USt-IdNr.: DE813185279 From Jon.Harald.Bovre at hafslund.no Thu Aug 19 12:27:39 2010 From: Jon.Harald.Bovre at hafslund.no (=?iso-8859-1?Q?B=F8vre_Jon_Harald?=) Date: Thu, 19 Aug 2010 12:27:39 +0200 Subject: SV: Google over IPv6 In-Reply-To: <181013.16512.qm@web55403.mail.re4.yahoo.com> References: <4183E1EF-97E2-451F-82B6-D050CE1EFD09@univie.ac.at> <181013.16512.qm@web55403.mail.re4.yahoo.com> Message-ID: Our network are not Google whitelisted, but for test we are still able to reach www.google.com over IPv6. You need to use a DNS from a whitelisted provider. I use Hurricane Electric 2001:470:20::2 as my primary DNS Works well, Youtube also start streaming in IPv6. But probably not very clever to use HE DNS on your DHCP toward customers. Jon Harald B?vre Hafslund Telekom Norway Fra: ipv6-ops-bounces+jon.harald.bovre=hafslund.no at lists.cluenet.de [mailto:ipv6-ops-bounces+jon.harald.bovre=hafslund.no at lists.cluenet.de] P? vegne av Matteo Bertozzi Sendt: 19. august 2010 12:10 Til: Ulrich Kiermayr Kopi: ipv6-ops at lists.cluenet.de Emne: Re: Google over IPv6 Hello Ulrich, I've tried http://ipv6.google.com and it works very well, even part of links inside the page are still on v4 site. Rgds Matteo ________________________________ From: Ulrich Kiermayr To: ipv6-ops at lists.cluenet.de Sent: Thu, August 19, 2010 11:22:55 AM Subject: Google over IPv6 Hello List, Since we were trying to reach Google regarding IPv6 (http://www.google.com/intl/en/ipv6/) for a while now without any success. Does anyone have an alternate way of reaching them, except what is mentioned on their webpage? lG uk -- Ulrich Kiermayr jabber xmpp:uk at jabber.univie.ac.at Leiter der Abteilung Datennetz und Telefonie skype:kiermayr Vienna University Computer Center phone +43 1 4277 14020 Universitaetsstrasse 7, 1010 Wien, AT fax +43 1 4277 9140 -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.cluenet.de/pipermail/ipv6-ops/attachments/20100819/3ae5f40e/attachment.html From ulrich.kiermayr at univie.ac.at Thu Aug 19 12:38:50 2010 From: ulrich.kiermayr at univie.ac.at (Ulrich Kiermayr) Date: Thu, 19 Aug 2010 12:38:50 +0200 Subject: SV: Google over IPv6 In-Reply-To: References: <4183E1EF-97E2-451F-82B6-D050CE1EFD09@univie.ac.at> <181013.16512.qm@web55403.mail.re4.yahoo.com> Message-ID: <4B46EF1E-90F7-4D39-AE4C-96A9CF548360@univie.ac.at> Hello, On 19.08.2010, at 12:27, B?vre Jon Harald wrote: > Our network are not Google whitelisted, but for test we are still able to reach www.google.com over IPv6. > > You need to use a DNS from a whitelisted provider. > I use Hurricane Electric 2001:470:20::2 as my primary DNS > Works well, Youtube also start streaming in IPv6. > But probably not very clever to use HE DNS on your DHCP toward customers. Then it would be as clever to simply override the entrys for www.google.com in our Nameservers and hardcode the AAAAs there. But this is not what we want. We are running all-dualstack for years now. and google is a way to generate v6 traffic. Esp. since all our Stundents in the WLAN networks do ipv6 (mostly without knowing/noticing). lG uk -- Ulrich Kiermayr jabber xmpp:uk at jabber.univie.ac.at Leiter der Abteilung Datennetz und Telefonie skype:kiermayr Vienna University Computer Center phone +43 1 4277 14020 Universitaetsstrasse 7, 1010 Wien, AT fax +43 1 4277 9140 From martin at millnert.se Thu Aug 19 12:45:10 2010 From: martin at millnert.se (Martin Millnert) Date: Thu, 19 Aug 2010 12:45:10 +0200 Subject: SV: Google over IPv6 In-Reply-To: <4B46EF1E-90F7-4D39-AE4C-96A9CF548360@univie.ac.at> References: <4183E1EF-97E2-451F-82B6-D050CE1EFD09@univie.ac.at> <181013.16512.qm@web55403.mail.re4.yahoo.com> <4B46EF1E-90F7-4D39-AE4C-96A9CF548360@univie.ac.at> Message-ID: <1282214710.20202.20.camel@hsa.vpn.anti> Hi Ulrich, On Thu, 2010-08-19 at 12:38 +0200, Ulrich Kiermayr wrote: > We are running all-dualstack for years now. and google is a way to > generate v6 traffic. Esp. since all our Stundents in the WLAN networks > do ipv6 (mostly without knowing/noticing). out of curiosity, (how) have you solved the rouge RA issue that ICS et al makes up? By what Steinar posted, it seems you **might** get away with setting up a separate set of resolvers that you can use in the WLAN network, if you commit to serving all v6 user experience issues with top service. This of course provided you have sufficient reachability to 15169. Regards, -- Martin Millnert -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 198 bytes Desc: This is a digitally signed message part Url : http://lists.cluenet.de/pipermail/ipv6-ops/attachments/20100819/5041c731/attachment.bin From ulrich.kiermayr at univie.ac.at Thu Aug 19 12:57:08 2010 From: ulrich.kiermayr at univie.ac.at (Ulrich Kiermayr) Date: Thu, 19 Aug 2010 12:57:08 +0200 Subject: SV: Google over IPv6 In-Reply-To: <1282214710.20202.20.camel@hsa.vpn.anti> References: <4183E1EF-97E2-451F-82B6-D050CE1EFD09@univie.ac.at> <181013.16512.qm@web55403.mail.re4.yahoo.com> <4B46EF1E-90F7-4D39-AE4C-96A9CF548360@univie.ac.at> <1282214710.20202.20.camel@hsa.vpn.anti> Message-ID: On 19.08.2010, at 12:45, Martin Millnert wrote: > Hi Ulrich, > > On Thu, 2010-08-19 at 12:38 +0200, Ulrich Kiermayr wrote: >> We are running all-dualstack for years now. and google is a way to >> generate v6 traffic. Esp. since all our Stundents in the WLAN networks >> do ipv6 (mostly without knowing/noticing). > > out of curiosity, (how) have you solved the rouge RA issue that ICS et > al makes up? We did 2 things. 1. Setting the RA Priority from our roouters to high (helps in most cases since winwows seems to respect this). 2. Monitoring the RAs in that network and inform/lock out users sending those RAs. In most cases the users simply were not aware, that they are sending this, they just turned on connection sharing "a long time ago and it never caused any trouble". And the support knows by now where to look at for a specific failure pattern (in our case: Internet works, but I cannot reach the webserver of the University nor can I send or receive email). What we also did is block IPv6 tunnel protocols at our border routers (we offer native ipv6 - so tunneling out of our lan is not a good idea anyway). All these measures help to almost eliminate the problem. But we are also looking for a more general solution to that. > By what Steinar posted, it seems you **might** get away with setting up > a separate set of resolvers that you can use in the WLAN network, if you > commit to serving all v6 user experience issues with top service. This > of course provided you have sufficient reachability to 15169. This should not be a problem here ;-) lG uk -- Ulrich Kiermayr jabber xmpp:uk at jabber.univie.ac.at Leiter der Abteilung Datennetz und Telefonie skype:kiermayr Vienna University Computer Center phone +43 1 4277 14020 Universitaetsstrasse 7, 1010 Wien, AT fax +43 1 4277 9140 From david.freedman at uk.clara.net Thu Aug 19 13:17:07 2010 From: david.freedman at uk.clara.net (David Freedman) Date: Thu, 19 Aug 2010 12:17:07 +0100 Subject: Google over IPv6 In-Reply-To: <4183E1EF-97E2-451F-82B6-D050CE1EFD09@univie.ac.at> References: <4183E1EF-97E2-451F-82B6-D050CE1EFD09@univie.ac.at> Message-ID: <4C6D12B3.9050104@uk.clara.net> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Ulrich Kiermayr wrote: > Hello List, > > Since we were trying to reach Google regarding IPv6 > (http://www.google.com/intl/en/ipv6/) for a while now without any > success. Does anyone have an alternate way of reaching them, except > what is mentioned on their webpage? > > lG uk This would be a great opportunity for Google to build a portal to allow ISPs to whitelist their own resolvers (perhaps after performing a number of automated tests to ensure compliance), I'm sure Google have a number of systems already which are designed for "ISP" accounts (i.e reputation/mail).... - -- David Freedman ClaraNET -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAkxtErMACgkQtFWeqpgEZrIdOQCgiGTYrDEppNFINs2TX3T1KCwP ULUAn3OJqDkljs/sPjeX95R5/sK5b0/+ =E5Pb -----END PGP SIGNATURE----- From ipv6-ops at ml.karotte.org Fri Aug 20 15:41:21 2010 From: ipv6-ops at ml.karotte.org (Sebastian Wiesinger) Date: Fri, 20 Aug 2010 15:41:21 +0200 Subject: Why you shouldn't worry about IPv6 just yet Message-ID: <20100820134121.GA31816@danton.fire-world.de> Wow, what a well researched and balanced article this is: http://www.pcpro.co.uk/realworld/360418/why-you-shouldnt-worry-about-ipv6-just-yet/ I'm pondering if this article is actually sarcasm or if he is trolling. Regards, Sebastian -- New GPG Key: 0x93A0B9CE (F4F6 B1A3 866B 26E9 450A 9D82 58A2 D94A 93A0 B9CE) Old GPG Key-ID: 0x76B79F20 (0x1B6034F476B79F20) 'Are you Death?' ... IT'S THE SCYTHE, ISN'T IT? PEOPLE ALWAYS NOTICE THE SCYTHE. -- Terry Pratchett, The Fifth Elephant From nanog at 85d5b20a518b8f6864949bd940457dc124746ddc.nosense.org Fri Aug 20 17:03:36 2010 From: nanog at 85d5b20a518b8f6864949bd940457dc124746ddc.nosense.org (Mark Smith) Date: Sat, 21 Aug 2010 00:33:36 +0930 Subject: Why you shouldn't worry about IPv6 just yet In-Reply-To: <20100820134121.GA31816@danton.fire-world.de> References: <20100820134121.GA31816@danton.fire-world.de> Message-ID: <20100821003336.705697ec@opy.nosense.org> On Fri, 20 Aug 2010 15:41:21 +0200 Sebastian Wiesinger wrote: > Wow, what a well researched and balanced article this is: > > http://www.pcpro.co.uk/realworld/360418/why-you-shouldnt-worry-about-ipv6-just-yet/ > > I'm pondering if this article is actually sarcasm or if he is > trolling. > So is he putting up is hand to be on call if and when something breaks? I'd much rather be too early than too late .... > Regards, > > Sebastian > > -- > New GPG Key: 0x93A0B9CE (F4F6 B1A3 866B 26E9 450A 9D82 58A2 D94A 93A0 B9CE) > Old GPG Key-ID: 0x76B79F20 (0x1B6034F476B79F20) > 'Are you Death?' ... IT'S THE SCYTHE, ISN'T IT? PEOPLE ALWAYS NOTICE THE SCYTHE. > -- Terry Pratchett, The Fifth Elephant From chris at timico.net Fri Aug 20 17:44:09 2010 From: chris at timico.net (Chris Nicholls) Date: Fri, 20 Aug 2010 16:44:09 +0100 Subject: IPv6 Message-ID: <20100820154409.GA48206@atsuko> Hello list, Anyone had experiance getting DNS glue records from SRSPlus? Ta -- Chris Nicholls Timico Network Operations chris at timico.net From michael at rancid.berkeley.edu Fri Aug 20 19:39:05 2010 From: michael at rancid.berkeley.edu (Michael Sinatra) Date: Fri, 20 Aug 2010 10:39:05 -0700 Subject: Why you shouldn't worry about IPv6 just yet In-Reply-To: <20100820134121.GA31816@danton.fire-world.de> References: <20100820134121.GA31816@danton.fire-world.de> Message-ID: <4C6EBDB9.6000409@rancid.berkeley.edu> On 08/20/10 06:41, Sebastian Wiesinger wrote: > Wow, what a well researched and balanced article this is: > > http://www.pcpro.co.uk/realworld/360418/why-you-shouldnt-worry-about-ipv6-just-yet/ > > I'm pondering if this article is actually sarcasm or if he is > trolling. > > Regards, > > Sebastian > Sorry, I stopped reading when I got to this line: "New devices are being added to the internet faster than Mexicans buy lottery tickets...." I am not one to cry "racism" all the time, but such a flip line seemed totally inappropriate, not to mention irrelevant. Such an article is not even worth reading. (And what I did read indeed seemed like baldfaced trolling, and he doesn't even appear to be terribly adept at that.) michael From ipv6-ops at ml.karotte.org Fri Aug 20 22:01:19 2010 From: ipv6-ops at ml.karotte.org (Sebastian Wiesinger) Date: Fri, 20 Aug 2010 22:01:19 +0200 Subject: Why you shouldn't worry about IPv6 just yet In-Reply-To: <4C6EBDB9.6000409@rancid.berkeley.edu> References: <20100820134121.GA31816@danton.fire-world.de> <4C6EBDB9.6000409@rancid.berkeley.edu> Message-ID: <20100820200119.GA2263@danton.fire-world.de> * Michael Sinatra [2010-08-20 19:40]: >> http://www.pcpro.co.uk/realworld/360418/why-you-shouldnt-worry-about-ipv6-just-yet/ >> >> I'm pondering if this article is actually sarcasm or if he is >> trolling. > > Sorry, I stopped reading when I got to this line: "New devices are being > added to the internet faster than Mexicans buy lottery tickets...." I > am not one to cry "racism" all the time, but such a flip line seemed > totally inappropriate, not to mention irrelevant. Such an article is Yes, that bothered me too. :( Regards, Sebastian -- New GPG Key: 0x93A0B9CE (F4F6 B1A3 866B 26E9 450A 9D82 58A2 D94A 93A0 B9CE) Old GPG Key-ID: 0x76B79F20 (0x1B6034F476B79F20) 'Are you Death?' ... IT'S THE SCYTHE, ISN'T IT? PEOPLE ALWAYS NOTICE THE SCYTHE. -- Terry Pratchett, The Fifth Elephant From tedm at ipinc.net Fri Aug 20 22:15:17 2010 From: tedm at ipinc.net (Ted Mittelstaedt) Date: Fri, 20 Aug 2010 13:15:17 -0700 Subject: Why you shouldn't worry about IPv6 just yet In-Reply-To: <20100820134121.GA31816@danton.fire-world.de> References: <20100820134121.GA31816@danton.fire-world.de> Message-ID: <4C6EE255.4050501@ipinc.net> On 8/20/2010 6:41 AM, Sebastian Wiesinger wrote: > Wow, what a well researched and balanced article this is: > > http://www.pcpro.co.uk/realworld/360418/why-you-shouldnt-worry-about-ipv6-just-yet/ > > I'm pondering if this article is actually sarcasm or if he is > trolling. > This is from the "The way I like it is the way it is. I got mine don't worry about his" school of thought. (with apologies to James Brown) Ted From brian.e.carpenter at gmail.com Fri Aug 20 22:45:45 2010 From: brian.e.carpenter at gmail.com (Brian E Carpenter) Date: Sat, 21 Aug 2010 08:45:45 +1200 Subject: Why you shouldn't worry about IPv6 just yet In-Reply-To: <4C6EE255.4050501@ipinc.net> References: <20100820134121.GA31816@danton.fire-world.de> <4C6EE255.4050501@ipinc.net> Message-ID: <4C6EE979.4030309@gmail.com> On 2010-08-21 08:15, Ted Mittelstaedt wrote: > > > On 8/20/2010 6:41 AM, Sebastian Wiesinger wrote: >> Wow, what a well researched and balanced article this is: >> >> http://www.pcpro.co.uk/realworld/360418/why-you-shouldnt-worry-about-ipv6-just-yet/ >> >> >> I'm pondering if this article is actually sarcasm or if he is >> trolling. >> > > This is from the "The way I like it is the way it is. I got mine don't > worry about his" school of thought. (with apologies to James Brown) Well, hold on. He isn't saying that ISPs shouldn't worry. He's saying that current ordinary users shouldn't worry. Is that so wrong? ISPs need to worry, content providers need to worry, and people who will be new Internet subscribers from about 2015 need to worry. Brian From tedm at ipinc.net Fri Aug 20 23:29:44 2010 From: tedm at ipinc.net (Ted Mittelstaedt) Date: Fri, 20 Aug 2010 14:29:44 -0700 Subject: Why you shouldn't worry about IPv6 just yet In-Reply-To: <4C6EE979.4030309@gmail.com> References: <20100820134121.GA31816@danton.fire-world.de> <4C6EE255.4050501@ipinc.net> <4C6EE979.4030309@gmail.com> Message-ID: <4C6EF3C8.1040307@ipinc.net> On 8/20/2010 1:45 PM, Brian E Carpenter wrote: > On 2010-08-21 08:15, Ted Mittelstaedt wrote: >> >> >> On 8/20/2010 6:41 AM, Sebastian Wiesinger wrote: >>> Wow, what a well researched and balanced article this is: >>> >>> http://www.pcpro.co.uk/realworld/360418/why-you-shouldnt-worry-about-ipv6-just-yet/ >>> >>> >>> I'm pondering if this article is actually sarcasm or if he is >>> trolling. >>> >> >> This is from the "The way I like it is the way it is. I got mine don't >> worry about his" school of thought. (with apologies to James Brown) > > Well, hold on. He isn't saying that ISPs shouldn't worry. He's saying that > current ordinary users shouldn't worry. Is that so wrong? > > ISPs need to worry, content providers need to worry, and people who will > be new Internet subscribers from about 2015 need to worry. > So if your an ordinary existing user in 2016 and your not worrying about IPv6 because you got yours, what do you do when you get laid off and find a new job in the next state - where your existing ISP isn't? With the churn rate in the business I think a lot of people will eventually fall into the classification of "new Internet subscribers" But more importantly, the new users getting on to the Internet are most likely the demographic without fixed buying habits. Meaning, the demographic that all companies want to advertise to. Advertisements from orgs are wasted on old farts like me who already have our IPv4 and who have been around to have established buying preferences. But they aren't wasted on young pups just getting online for the first time who have no established preferences. Those young pups are just getting computers for the first time, just getting service for the first time, etc. They will be on IPv6. And advertisers will want to go to them. And content providers paid by those advertisers will want to go to them. So they will go to IPv6 since that's where those target markets will be at. You can fence yourself in but you can never fence the world out. Ted > Brian From ms at man-da.de Fri Aug 20 23:37:59 2010 From: ms at man-da.de (Marcus Stoegbauer) Date: Fri, 20 Aug 2010 23:37:59 +0200 Subject: Why you shouldn't worry about IPv6 just yet In-Reply-To: <20100821003336.705697ec@opy.nosense.org> References: <20100820134121.GA31816@danton.fire-world.de> <20100821003336.705697ec@opy.nosense.org> Message-ID: <4C6EF5B7.2070500@man-da.de> On 20.08.10 17:03, Mark Smith wrote: > On Fri, 20 Aug 2010 15:41:21 +0200 > Sebastian Wiesinger wrote: > >> Wow, what a well researched and balanced article this is: >> >> http://www.pcpro.co.uk/realworld/360418/why-you-shouldnt-worry-about-ipv6-just-yet/ >> >> I'm pondering if this article is actually sarcasm or if he is >> trolling. >> > > So is he putting up is hand to be on call if and when something breaks? > > I'd much rather be too early than too late .... That's what she said. Marcus But on a serious note: The article shows the viewpoint of internet users. In theory they need not worry because they pay their providers for presenting them with a solution. In practice the article barely makes this distinction because it would've distracted from the authors imaginary wit. From brian.e.carpenter at gmail.com Sat Aug 21 00:10:10 2010 From: brian.e.carpenter at gmail.com (Brian E Carpenter) Date: Sat, 21 Aug 2010 10:10:10 +1200 Subject: Why you shouldn't worry about IPv6 just yet In-Reply-To: <4C6EF3C8.1040307@ipinc.net> References: <20100820134121.GA31816@danton.fire-world.de> <4C6EE255.4050501@ipinc.net> <4C6EE979.4030309@gmail.com> <4C6EF3C8.1040307@ipinc.net> Message-ID: <4C6EFD42.4090006@gmail.com> Ted, On 2010-08-21 09:29, Ted Mittelstaedt wrote: > > > On 8/20/2010 1:45 PM, Brian E Carpenter wrote: >> On 2010-08-21 08:15, Ted Mittelstaedt wrote: >>> >>> >>> On 8/20/2010 6:41 AM, Sebastian Wiesinger wrote: >>>> Wow, what a well researched and balanced article this is: >>>> >>>> http://www.pcpro.co.uk/realworld/360418/why-you-shouldnt-worry-about-ipv6-just-yet/ >>>> >>>> >>>> >>>> I'm pondering if this article is actually sarcasm or if he is >>>> trolling. >>>> >>> >>> This is from the "The way I like it is the way it is. I got mine don't >>> worry about his" school of thought. (with apologies to James Brown) >> >> Well, hold on. He isn't saying that ISPs shouldn't worry. He's saying >> that >> current ordinary users shouldn't worry. Is that so wrong? >> >> ISPs need to worry, content providers need to worry, and people who will >> be new Internet subscribers from about 2015 need to worry. >> > > So if your an ordinary existing user in 2016 and your not worrying > about IPv6 because you got yours, what do you do when you get laid > off and find a new job in the next state - where your existing ISP > isn't? True. But then the new ISP won't get the business, and its competitor who has v6 will. At least, that's the theory. So ISPs should worry... Brian > > With the churn rate in the business I think a lot of people will > eventually fall into the classification of "new Internet subscribers" > > But more importantly, the new users getting on to the Internet are > most likely the demographic without fixed buying habits. Meaning, the > demographic that all companies want to advertise to. > > Advertisements from orgs are wasted on old farts like me who already > have our IPv4 and who have been around to have established buying > preferences. > > But they aren't wasted on young pups just getting online for the first > time who have no established preferences. Those young pups are just > getting computers for the first time, just getting service for the first > time, etc. They will be on IPv6. And advertisers will want to go to > them. And content providers paid by those advertisers will want to go > to them. So they will go to IPv6 since that's where those target > markets will be at. > > You can fence yourself in but you can never fence the world out. > > Ted > >> Brian > From tedm at ipinc.net Sat Aug 21 00:52:50 2010 From: tedm at ipinc.net (Ted Mittelstaedt) Date: Fri, 20 Aug 2010 15:52:50 -0700 Subject: Why you shouldn't worry about IPv6 just yet In-Reply-To: <4C6EFD42.4090006@gmail.com> References: <20100820134121.GA31816@danton.fire-world.de> <4C6EE255.4050501@ipinc.net> <4C6EE979.4030309@gmail.com> <4C6EF3C8.1040307@ipinc.net> <4C6EFD42.4090006@gmail.com> Message-ID: <4C6F0742.40306@ipinc.net> On 8/20/2010 3:10 PM, Brian E Carpenter wrote: > Ted, > > On 2010-08-21 09:29, Ted Mittelstaedt wrote: >> >> >> On 8/20/2010 1:45 PM, Brian E Carpenter wrote: >>> On 2010-08-21 08:15, Ted Mittelstaedt wrote: >>>> >>>> >>>> On 8/20/2010 6:41 AM, Sebastian Wiesinger wrote: >>>>> Wow, what a well researched and balanced article this is: >>>>> >>>>> http://www.pcpro.co.uk/realworld/360418/why-you-shouldnt-worry-about-ipv6-just-yet/ >>>>> >>>>> >>>>> >>>>> I'm pondering if this article is actually sarcasm or if he is >>>>> trolling. >>>>> >>>> >>>> This is from the "The way I like it is the way it is. I got mine don't >>>> worry about his" school of thought. (with apologies to James Brown) >>> >>> Well, hold on. He isn't saying that ISPs shouldn't worry. He's saying >>> that >>> current ordinary users shouldn't worry. Is that so wrong? >>> >>> ISPs need to worry, content providers need to worry, and people who will >>> be new Internet subscribers from about 2015 need to worry. >>> >> >> So if your an ordinary existing user in 2016 and your not worrying >> about IPv6 because you got yours, what do you do when you get laid >> off and find a new job in the next state - where your existing ISP >> isn't? > I guess I should avoid ending sentences with prepositions that I've converted into verbs. where your existing ISP isn't -located-? And you only GET IPv6 from the new ISP I think that would point the indication to that users DO have to worry about IPv6. Ted > True. But then the new ISP won't get the business, and its competitor > who has v6 will. At least, that's the theory. So ISPs should worry... > > Brian >> >> With the churn rate in the business I think a lot of people will >> eventually fall into the classification of "new Internet subscribers" >> >> But more importantly, the new users getting on to the Internet are >> most likely the demographic without fixed buying habits. Meaning, the >> demographic that all companies want to advertise to. >> >> Advertisements from orgs are wasted on old farts like me who already >> have our IPv4 and who have been around to have established buying >> preferences. >> >> But they aren't wasted on young pups just getting online for the first >> time who have no established preferences. Those young pups are just >> getting computers for the first time, just getting service for the first >> time, etc. They will be on IPv6. And advertisers will want to go to >> them. And content providers paid by those advertisers will want to go >> to them. So they will go to IPv6 since that's where those target >> markets will be at. >> >> You can fence yourself in but you can never fence the world out. >> >> Ted >> >>> Brian >> From dougb at dougbarton.us Sat Aug 21 01:38:04 2010 From: dougb at dougbarton.us (Doug Barton) Date: Fri, 20 Aug 2010 16:38:04 -0700 Subject: Why you shouldn't worry about IPv6 just yet In-Reply-To: <4C6EE979.4030309@gmail.com> References: <20100820134121.GA31816@danton.fire-world.de> <4C6EE255.4050501@ipinc.net> <4C6EE979.4030309@gmail.com> Message-ID: <4C6F11DC.7020509@dougbarton.us> On 08/20/2010 13:45, Brian E Carpenter wrote: > Well, hold on. He isn't saying that ISPs shouldn't worry. He's saying that > current ordinary users shouldn't worry. Is that so wrong? > > ISPs need to worry, content providers need to worry, and people who will > be new Internet subscribers from about 2015 need to worry. Aside from the fact that (as others have already pointed out) most of the article is written in very bad taste; my impression was sort of the same as Brian's. *We* need to focus on the small details, but users don't want this "eye pee vee six" thing, they want web, mail, movies, etc. The details we need to focus on are how to make that work. Doug -- Improve the effectiveness of your Internet presence with a domain name makeover! http://SupersetSolutions.com/ Computers are useless. They can only give you answers. -- Pablo Picasso From merike at doubleshotsecurity.com Sat Aug 21 04:20:58 2010 From: merike at doubleshotsecurity.com (Merike Kaeo) Date: Fri, 20 Aug 2010 19:20:58 -0700 Subject: Why you shouldn't worry about IPv6 just yet In-Reply-To: <4C6F11DC.7020509@dougbarton.us> References: <20100820134121.GA31816@danton.fire-world.de> <4C6EE255.4050501@ipinc.net> <4C6EE979.4030309@gmail.com> <4C6F11DC.7020509@dougbarton.us> Message-ID: On Aug 20, 2010, at 4:38 PM, Doug Barton wrote: > On 08/20/2010 13:45, Brian E Carpenter wrote: > >> Well, hold on. He isn't saying that ISPs shouldn't worry. He's saying that >> current ordinary users shouldn't worry. Is that so wrong? >> >> ISPs need to worry, content providers need to worry, and people who will >> be new Internet subscribers from about 2015 need to worry. > > Aside from the fact that (as others have already pointed out) most of the article is written in very bad taste; my impression was sort of the same as Brian's. *We* need to focus on the small details, but users don't want this "eye pee vee six" thing, they want web, mail, movies, etc. The details we need to focus on are how to make that work. > My take is he's writing from end user's point of view as others have stated but I was left with a clear impression that this guy has little understanding of networking. Scary if he's an app writer for scify channel. As if win7 and Leopard are first OSs to ship native v6 default.....where's he been last 4-5 years?!? I do agree that end users will not care as to IP protocols but applications developers should so they don't write code that locks them into doing non-optimal non-forward thinking solutions. I'm not an app coder but have heard stories of folks running into issues by assuming 32 bit addresses....probably other nuances exist Am sure google and facebook have lots of great stories here that they'll never divulge. Everything is 'easy' once you spent the zillions of hours debugging, optimizing and making things work. There is a learning curve and v6 is getting deployed (we all know this). Dan Wing has an interesting draft: http://tools.ietf.org/html/draft-wing-http-new-tech-00 What I worry about is ntp, syslog, radius/tacacs+, snmp, etc etc running natively over v6 transport. Sure, v4 in a dual stack environment fine for now but for how long?!? Some v6 solutions exist, not many. [I'm running a testbed right now to test some of this.....includes cisco, juniper, OSX, Ubuntu, win7, maybe 2008 server, and some CPE devices]. I'll make public what I find when I finish. NAT gives me nightmares... - merike From spz at serpens.de Sat Aug 21 07:28:15 2010 From: spz at serpens.de (S.P.Zeidler) Date: Sat, 21 Aug 2010 07:28:15 +0200 Subject: Why you shouldn't worry about IPv6 just yet In-Reply-To: <4C6F0742.40306@ipinc.net> References: <20100820134121.GA31816@danton.fire-world.de> <4C6EE255.4050501@ipinc.net> <4C6EE979.4030309@gmail.com> <4C6EF3C8.1040307@ipinc.net> <4C6EFD42.4090006@gmail.com> <4C6F0742.40306@ipinc.net> Message-ID: <20100821052814.GU698@serpens.de> Hi, Thus wrote Ted Mittelstaedt (tedm at ipinc.net): > I guess I should avoid ending sentences with prepositions that I've > converted into verbs. > > where your existing ISP isn't -located-? > > And you only GET IPv6 from the new ISP Frankly, for a true end user who is moving, packing the kitchen will be more work than buying a new, IPv6 capable "DSL-router" (if they don't get one from their provider and their old one doesn't accidentially support it) and getting it going with the cheat sheet their provider handed to them. > I think that would point the indication to that users DO have to > worry about IPv6. You expect that end users have more knowledge of IPv4 than what can be replaced in 30min? :) -They- really can cross that brook when they come to it. The need to plan is where replacing kit is an investment, and the serious work involved is also not in most home networks (yours may differ, but people for whom that is true are likely to be running IPv6 already). regards, spz -- spz at serpens.de (S.P.Zeidler) From bmanning at vacation.karoshi.com Sat Aug 21 09:06:00 2010 From: bmanning at vacation.karoshi.com (bmanning at vacation.karoshi.com) Date: Sat, 21 Aug 2010 07:06:00 +0000 Subject: Why you shouldn't worry about IPv6 just yet In-Reply-To: <4C6EF5B7.2070500@man-da.de> References: <20100820134121.GA31816@danton.fire-world.de> <20100821003336.705697ec@opy.nosense.org> <4C6EF5B7.2070500@man-da.de> Message-ID: <20100821070600.GB1654@vacation.karoshi.com.> > > > > I'd much rather be too early than too late .... > > That's what she said. > > Marcus > > But on a serious note: The article shows the viewpoint of internet > users. In theory they need not worry because they pay their providers > for presenting them with a solution. actually, he is making the argument for NAT being a fundamental technlogy for the duration... and so as long as the ISP provides the "outside" functionality, he can be happy running 192.168.18.0/24 inside for the next x0 years. and having been early (anyone remember the prefix before 3ffe::?) a number of times, this time i'd rather be late. but i'm not. its possible to "one-up" the game and convert to using all (native) IPv6 for your internal network and on the other side of your NAT have the IPv4 stuff - until your ISP gets its act together. I've been using the IVI code (as described at SIGCOM 2007) to run native IPv6 on the house network and IPv4 on the outside... you'll not see my IPv6 prefix -unless- you have native IPv6 transit to me. if you don't like the crusty IVI code, I've heard there is something else outhere called AFTR? which might do the same thing. --bill From lists at quux.de Sat Aug 21 09:35:48 2010 From: lists at quux.de (Jens Link) Date: Sat, 21 Aug 2010 09:35:48 +0200 Subject: Why you shouldn't worry about IPv6 just yet In-Reply-To: <4C6EF5B7.2070500@man-da.de> (Marcus Stoegbauer's message of "Fri, 20 Aug 2010 23:37:59 +0200") References: <20100820134121.GA31816@danton.fire-world.de> <20100821003336.705697ec@opy.nosense.org> <4C6EF5B7.2070500@man-da.de> Message-ID: <87aaog8lkb.fsf@oban.berlin.quux.de> Marcus Stoegbauer writes: > But on a serious note: The article shows the viewpoint of internet > users. The author of this article is a "networks expert". Well i guess every body who is able to setup his home DSL router is an expert. Jens -- ------------------------------------------------------------------------- | Foelderichstr. 40 | 13595 Berlin, Germany | +49-151-18721264 | | http://blog.quux.de | jabber: jenslink at guug.de | ------------------- | ------------------------------------------------------------------------- From tedm at ipinc.net Sat Aug 21 15:54:00 2010 From: tedm at ipinc.net (Ted Mittelstaedt) Date: Sat, 21 Aug 2010 06:54:00 -0700 Subject: Why you shouldn't worry about IPv6 just yet In-Reply-To: <20100821052814.GU698@serpens.de> References: <20100820134121.GA31816@danton.fire-world.de> <4C6EE255.4050501@ipinc.net> <4C6EE979.4030309@gmail.com> <4C6EF3C8.1040307@ipinc.net> <4C6EFD42.4090006@gmail.com> <4C6F0742.40306@ipinc.net> <20100821052814.GU698@serpens.de> Message-ID: <4C6FDA78.9090902@ipinc.net> On 8/20/2010 10:28 PM, S.P.Zeidler wrote: > Hi, > > Thus wrote Ted Mittelstaedt (tedm at ipinc.net): > >> I guess I should avoid ending sentences with prepositions that I've >> converted into verbs. >> >> where your existing ISP isn't -located-? >> >> And you only GET IPv6 from the new ISP > > Frankly, for a true end user who is moving, packing the kitchen will be > more work than buying a new, IPv6 capable "DSL-router" (if they don't get > one from their provider and their old one doesn't accidentially support > it) and getting it going with the cheat sheet their provider handed to > them. > I think the article was written for the end user. Generally you can assume that anytime an author's credentials consist of "30 years in the industry" that's publishing-speak for "aimed at end users" Ted >> I think that would point the indication to that users DO have to >> worry about IPv6. > > You expect that end users have more knowledge of IPv4 than what can be > replaced in 30min? :) -They- really can cross that brook when they come > to it. > The need to plan is where replacing kit is an investment, and the serious > work involved is also not in most home networks (yours may differ, but > people for whom that is true are likely to be running IPv6 already). > > regards, > spz From fred at cisco.com Sun Aug 22 05:09:45 2010 From: fred at cisco.com (Fred Baker) Date: Sat, 21 Aug 2010 20:09:45 -0700 Subject: Why you shouldn't worry about IPv6 just yet In-Reply-To: <20100821070600.GB1654@vacation.karoshi.com.> References: <20100820134121.GA31816@danton.fire-world.de> <20100821003336.705697ec@opy.nosense.org> <4C6EF5B7.2070500@man-da.de> <20100821070600.GB1654@vacation.karoshi.com.> Message-ID: On Aug 21, 2010, at 12:06 AM, bmanning at vacation.karoshi.com wrote: > if you don't like the crusty IVI code, I've heard there is > something else outhere called AFTR? which might do the same thing. It doesn't. It runs DS-lite. DS-lite assumes you have an IPv6 network and enables you to tunnel your IPv4 over it. Upstream, there might be a traditional IPv4/IPv4 NAT. So it gives you native IPv6 connectivity and dual-NAT'd tunneled IPv4 connectivity. The premise is that you have a part of your network - at least one router or at least one link - that you have reconfigured as IPv6-only, and that the periphery of that ipv6-only domain is populated with some number of these things. As your IPv6-only domain grows, the distance between the tunnel endpoints grows. The premise is that your network is, within that domain at least, primarily IPv6, but the service is primarily IPv4. The downside of it is that there always has to an an "other end" for your IPv4 service, and it has to be in a provider you trust. In 6rd, Free.fr did the opposite. The started with their IPv4 infrastructure, and provided an application that (because they provisioned it that way) only ran on IPv6, and came up with an IPv6/IPv4 tunneled infrastructure that could provide that service. The service turns out to be popular, something akin to youtube. So they are giving you what appears to be a dual stack infrastructure, but is in fact IPv4 plus IPv6/IPv4. As they decide they want to, they can of course change to dual stack as they choose, obsoleting the tunneled bit, and when it suits them they can turn IPv4 off. So for 6rd, the premise is that the service is applications, at least one of them runs IPv6 (which happens to run on IPv4 but can be changed to native service without the user knowing) and at least one of them runs on IPv4. unlike ds-lite, which is trying to help you build an ipv6-only infrastructure, it tries to help you build a dual stack infrastructure. Translation, which IVI is part of, starts from the assumption that some of your systems are ipv4-only and some are ipv6-only, and they need to talk with each other. translation accomplishes that, to at least some extent. From kiwi at oav.net Tue Aug 24 10:05:02 2010 From: kiwi at oav.net (Xavier Beaudouin) Date: Tue, 24 Aug 2010 10:05:02 +0200 Subject: Why you shouldn't worry about IPv6 just yet In-Reply-To: <4C6EBDB9.6000409@rancid.berkeley.edu> References: <20100820134121.GA31816@danton.fire-world.de> <4C6EBDB9.6000409@rancid.berkeley.edu> Message-ID: Hi There, Le 20 ao?t 2010 ? 19:39, Michael Sinatra a ?crit : > On 08/20/10 06:41, Sebastian Wiesinger wrote: >> Wow, what a well researched and balanced article this is: >> >> http://www.pcpro.co.uk/realworld/360418/why-you-shouldnt-worry-about-ipv6-just-yet/ >> >> I'm pondering if this article is actually sarcasm or if he is >> trolling. >> >> Regards, >> >> Sebastian >> > > Sorry, I stopped reading when I got to this line: "New devices are being added to the internet faster than Mexicans buy lottery tickets...." I am not one to cry "racism" all the time, but such a flip line seemed totally inappropriate, not to mention irrelevant. Such an article is not even worth reading. (And what I did read indeed seemed like baldfaced trolling, and he doesn't even appear to be terribly adept at that.) It sounds more sarcasm and trolling than really good journalism work. We (all IPv6 guys there) will see how he will handle the lack of IPv4 and the needs to change all of his routers at home. Just don't rely too much such bullshits... We are in a time that we still have time to tests, and doing experiments about IPv6. Some people are not deploying and gets problems and reply to good questions about how to make this works... it is more or less like evolutions when dinosaurs desapears from earth.... :) My O,02?. /Xavier From jjmbcom at gmail.com Tue Aug 24 14:19:13 2010 From: jjmbcom at gmail.com (John Jason Brzozowski) Date: Tue, 24 Aug 2010 08:19:13 -0400 Subject: Why you shouldn't worry about IPv6 just yet In-Reply-To: <4C6EBDB9.6000409@rancid.berkeley.edu> References: <20100820134121.GA31816@danton.fire-world.de> <4C6EBDB9.6000409@rancid.berkeley.edu> Message-ID: This article somehow managed to make it around my internal list at work. We stopped reading shortly after the text below too. It was hard to take this article seriously. John On Fri, Aug 20, 2010 at 1:39 PM, Michael Sinatra < michael at rancid.berkeley.edu> wrote: > On 08/20/10 06:41, Sebastian Wiesinger wrote: > >> Wow, what a well researched and balanced article this is: >> >> >> http://www.pcpro.co.uk/realworld/360418/why-you-shouldnt-worry-about-ipv6-just-yet/ >> >> I'm pondering if this article is actually sarcasm or if he is >> trolling. >> >> Regards, >> >> Sebastian >> >> > Sorry, I stopped reading when I got to this line: "New devices are being > added to the internet faster than Mexicans buy lottery tickets...." I am > not one to cry "racism" all the time, but such a flip line seemed totally > inappropriate, not to mention irrelevant. Such an article is not even worth > reading. (And what I did read indeed seemed like baldfaced trolling, and he > doesn't even appear to be terribly adept at that.) > > michael > -- =================================== John Jason Brzozowski =================================== -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.cluenet.de/pipermail/ipv6-ops/attachments/20100824/cf094d38/attachment.html From marcoh at marcoh.net Thu Aug 26 17:22:34 2010 From: marcoh at marcoh.net (Marco Hogewoning) Date: Thu, 26 Aug 2010 17:22:34 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers Message-ID: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> Sorry for the spam... Groet, MarcoH --- PRESS RELEASE 26 August 2010 IPv6 available for all, starting today XS4ALL OFFERS IPv6 TO ALL CUSTOMERS Dutch ISP XS4ALL has officially started providing IPv6 connectivity to all its customers today. The current pool of available IP addresses (IPv4) will likely run dry within one year, forcing ISPs to take action. When there are no longer IPv4 addresses available, new services can only be connected to the Internet via IPv6. XS4ALL has been working hard on implementing the new protocol and the hard labour is paying off: XS4ALL is among the very first providers in Europe to make IPv6 available to its customers, offering them the chance to prepare for the future. Successful pilot XS4ALL customers have been able to experiment with IPv6 for some time now, using so called IPv6 tunnelling. This tunnelling technique provides IPv6 connectivity, but is still dependent upon the underlying IPv4 infrastructure. Today?s new technology however offers native IPv6, totally independent of the old IPv4 protocol, which makes it truly future proof. To be able to provide large scale IPv6 connectivity, XS4ALL has recently completed a very successful pilot in which several hundred customers have been testing the new protocol. Thanks to their co-operation and feedback XS4ALL was able to identify several minor problems in its network. With those fixed, we feel confident that the protocol can now be released to the general public. What?s in a name ?Seventeen years ago we were the first to provide internet access to the public in The Netherlands,? says Niels Huijbregts, XS4ALL?s spokesperson. ?Now we?re the first to ensure future connectivity as the Internet expands to cover more and more devices and places. Once again we?re providing Access for All?. IP addresses are the unique numbers that identify every computer on the Internet. Because of the rapid growth of the net, new addresses are running out. That?s why it?s of major importance to start using the new protocol: when all IPv4 addresses have been given out, new services can only be connected to the Internet using IPv6. ?If you start using the new protocol now, there?s still time to get used to it and experiment. If you wait too long you run the risk of having to switch in great haste,? says Huijbregts. Flick of a switch XS4ALL customers can activate IPv6 on their account with a simple flick of a switch in the Service Centre on the company?s website. They can start using the new protocol right away, provided they have DSL modem that supports IPv6. Suitable new modems will be provided to customers who renew their contract. Now that this important milestone has been reached, XS4ALL will focus on adjusting all mail and web servers, the last step in making the company fully IPv6-ready. For more information please call Niels Huijbregts, XS4ALL?s spokesperson From gert at space.net Thu Aug 26 17:30:33 2010 From: gert at space.net (Gert Doering) Date: Thu, 26 Aug 2010 17:30:33 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> Message-ID: <20100826153033.GA61734@Space.Net> Hi, On Thu, Aug 26, 2010 at 05:22:34PM +0200, Marco Hogewoning wrote: > Sorry for the spam... Thanks for the information! And congratulations for going ahead! Gert Doering -- NetMaster -- Total number of prefixes smaller than registry allocations: 155817 SpaceNet AG Vorstand: Sebastian v. Bomhard Joseph-Dollinger-Bogen 14 Aufsichtsratsvors.: A. Grundner-Culemann D-80807 Muenchen HRB: 136055 (AG Muenchen) Tel: +49 (89) 32356-444 USt-IdNr.: DE813185279 From M.Hotze at hotze.com Thu Aug 26 17:31:32 2010 From: M.Hotze at hotze.com (Martin Hotze) Date: Thu, 26 Aug 2010 15:31:32 +0000 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> Message-ID: <6A4EBE06CF13034585C3F773EAF836A32F492A25@exsrv01.hotzecom.local> > -----Original Message----- > From: ipv6-ops-bounces+m.hotze=hotze.com at lists.cluenet.de [mailto:ipv6- > ops-bounces+m.hotze=hotze.com at lists.cluenet.de] On Behalf Of Marco > Hogewoning > Sent: Thursday, August 26, 2010 5:23 PM > To: ipv6-ops at lists.cluenet.de > Subject: XS4ALL Introduces native IPv6 for DSL customers > (...) > XS4ALL OFFERS IPv6 TO ALL CUSTOMERS congrats. (...) > XS4ALL customers can activate IPv6 on their account with a simple flick of > a switch in the Service Centre on the company's website. They can start > using the new protocol right away, provided they have DSL modem that > supports IPv6. Suitable new modems will be provided to customers who renew > their contract. So: what modems do you use or what hardware do you suggest to your customers? #m From david.freedman at uk.clara.net Thu Aug 26 17:39:23 2010 From: david.freedman at uk.clara.net (David Freedman) Date: Thu, 26 Aug 2010 16:39:23 +0100 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> Message-ID: <4C768AAB.5090004@uk.clara.net> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 > PRESS RELEASE 26 August 2010 > > IPv6 available for all, starting today > > XS4ALL OFFERS IPv6 TO ALL CUSTOMERS Congrats, Gone with Fritzbox in the end for non-cisco xDSL then? Dave. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAkx2iqsACgkQtFWeqpgEZrIexQCZAUisWAR1qGC/MyYv/PrV+2AR rYMAoMJRM7+3U2XvbiFKOeVWwwHaPlYl =8h4y -----END PGP SIGNATURE----- From daniel at bit.nl Thu Aug 26 17:43:06 2010 From: daniel at bit.nl (Daniel Verlouw) Date: Thu, 26 Aug 2010 17:43:06 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> Message-ID: <1282837386.31961.21.camel@daniel> Excellent news, good job! On Thu, 2010-08-26 at 17:22 +0200, Marco Hogewoning wrote: > XS4ALL customers can activate IPv6 on their account with a simple flick of a switch in the Service Centre on the company?s website. just out of curiosity why not turned on by default ? Router/modem compatibility issues? A provisioning thing? Or afraid of helpdesk overload with calls such as "my connection all of sudden doesn't work anymore after you turned on this Eye Pee Vee Six thing" ? ;) --Daniel. From d.bay at rrbone-bb.net Thu Aug 26 18:29:34 2010 From: d.bay at rrbone-bb.net (Dominik Bay) Date: Thu, 26 Aug 2010 18:29:34 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <1282837386.31961.21.camel@daniel> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <1282837386.31961.21.camel@daniel> Message-ID: <20100826182934.1d85b78f@libelle.kk18.dtm.openap.net> On Thu, 26 Aug 2010 17:43:06 +0200 Daniel Verlouw wrote: > Excellent news, good job! > > On Thu, 2010-08-26 at 17:22 +0200, Marco Hogewoning wrote: > > XS4ALL customers can activate IPv6 on their account with a simple > > flick of a switch in the Service Centre on the company?s website. > > just out of curiosity why not turned on by default ? Router/modem > compatibility issues? A provisioning thing? Or afraid of helpdesk > overload with calls such as "my connection all of sudden doesn't work > anymore after you turned on this Eye Pee Vee Six thing" ? ;) Just as a side node on the provisioning-part: We are currently testing DOCSIS Cablemodems and CM with integrated Routers. This hopefully brings us soon the ability to detect and disable IPv6 for known devices and software versions which break IPv6 connectivity. I would then turn it on by default for devices which are known to be working. I *guess* that this could be an approach for TR069 enabled DSL devices, just checking the software version, and setting a RADIUS Flag for enabling IPCP6. Anyone having such checks implemented yet? Kind regards, Dominik From marcoh at marcoh.net Thu Aug 26 18:51:05 2010 From: marcoh at marcoh.net (Marco Hogewoning) Date: Thu, 26 Aug 2010 18:51:05 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <6A4EBE06CF13034585C3F773EAF836A32F492A25@exsrv01.hotzecom.local> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <6A4EBE06CF13034585C3F773EAF836A32F492A25@exsrv01.hotzecom.local> Message-ID: > So: what modems do you use or what hardware do you suggest to your customers? A stripped down version of AVM Fritz!Boz 7390, called the 7340. IPv6 firmware for it is currently in beta, expected release is autumn this year. But it's available at the moment, but no guarantees. Groet, MarcoH From marcoh at marcoh.net Thu Aug 26 18:52:37 2010 From: marcoh at marcoh.net (Marco Hogewoning) Date: Thu, 26 Aug 2010 18:52:37 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <1282837386.31961.21.camel@daniel> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <1282837386.31961.21.camel@daniel> Message-ID: <322E2633-C041-428F-92F3-25F38DAE6F5D@marcoh.net> On 26 aug 2010, at 17:43, Daniel Verlouw wrote: > Excellent news, good job! > > On Thu, 2010-08-26 at 17:22 +0200, Marco Hogewoning wrote: >> XS4ALL customers can activate IPv6 on their account with a simple flick of a switch in the Service Centre on the company?s website. > > just out of curiosity why not turned on by default ? Router/modem > compatibility issues? A provisioning thing? Or afraid of helpdesk > overload with calls such as "my connection all of sudden doesn't work > anymore after you turned on this Eye Pee Vee Six thing" ? ;) Both, on one side we know we have a lot of non-IPv6 capable modems out there, so why waist resources on them. The other is indeed the fact taht if we push IPv6 to people without them being aware we might run into trouble. Groet, MarcoH From dr at cluenet.de Thu Aug 26 21:11:50 2010 From: dr at cluenet.de (Daniel Roesen) Date: Thu, 26 Aug 2010 21:11:50 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <6A4EBE06CF13034585C3F773EAF836A32F492A25@exsrv01.hotzecom.local> Message-ID: <20100826191150.GA10681@srv03.cluenet.de> On Thu, Aug 26, 2010 at 06:51:05PM +0200, Marco Hogewoning wrote: > > So: what modems do you use or what hardware do you suggest to your customers? > > A stripped down version of AVM Fritz!Boz 7390, called the 7340. Interestingly, this german artical says 7270+7570: http://www.zdnet.de/news/wirtschaft_telekommunikation_avm_und_xs4all_bieten_natives_ipv6_protokoll_fuer_privatanwender_story-39001023-41531444-1.htm Congrats btw Best regards, Daniel -- CLUE-RIPE -- Jabber: dr at cluenet.de -- dr at IRCnet -- PGP: 0xA85C8AA0 From dr at cluenet.de Thu Aug 26 21:13:41 2010 From: dr at cluenet.de (Daniel Roesen) Date: Thu, 26 Aug 2010 21:13:41 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <322E2633-C041-428F-92F3-25F38DAE6F5D@marcoh.net> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <1282837386.31961.21.camel@daniel> <322E2633-C041-428F-92F3-25F38DAE6F5D@marcoh.net> Message-ID: <20100826191341.GB10681@srv03.cluenet.de> On Thu, Aug 26, 2010 at 06:52:37PM +0200, Marco Hogewoning wrote: > Both, on one side we know we have a lot of non-IPv6 capable modems > out there, Modems, or actually routers (with built-in modems)? I guess you do DHCP-PD for customer's LAN interface? What size of LAN block does a customer get, just a /64? Or /56 and first /64 out of that getting assigned to the LAN interface? Best regards, Daniel -- CLUE-RIPE -- Jabber: dr at cluenet.de -- dr at IRCnet -- PGP: 0xA85C8AA0 From nanog at 85d5b20a518b8f6864949bd940457dc124746ddc.nosense.org Thu Aug 26 23:53:44 2010 From: nanog at 85d5b20a518b8f6864949bd940457dc124746ddc.nosense.org (Mark Smith) Date: Fri, 27 Aug 2010 07:23:44 +0930 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <322E2633-C041-428F-92F3-25F38DAE6F5D@marcoh.net> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <1282837386.31961.21.camel@daniel> <322E2633-C041-428F-92F3-25F38DAE6F5D@marcoh.net> Message-ID: <20100827072344.636bf9b5@opy.nosense.org> On Thu, 26 Aug 2010 18:52:37 +0200 Marco Hogewoning wrote: > > On 26 aug 2010, at 17:43, Daniel Verlouw wrote: > > > Excellent news, good job! > > > > On Thu, 2010-08-26 at 17:22 +0200, Marco Hogewoning wrote: > >> XS4ALL customers can activate IPv6 on their account with a simple flick of a switch in the Service Centre on the company?s website. > > > > just out of curiosity why not turned on by default ? Router/modem > > compatibility issues? A provisioning thing? Or afraid of helpdesk > > overload with calls such as "my connection all of sudden doesn't work > > anymore after you turned on this Eye Pee Vee Six thing" ? ;) > > > Both, on one side we know we have a lot of non-IPv6 capable modems out there, so why waist resources on them. The other is indeed the fact taht if we push IPv6 to people without them being aware we might run into trouble. > I've seen an IPv4 only modem lock up two minutes after establishing a PPP session, and then having upstream LNS periodically try to kick off IPV6CP again. I suppose a lock up is somewhat equivalent to a PPP Protocol-Reject ;-) Regards, Mark. From marcoh at marcoh.net Fri Aug 27 00:20:27 2010 From: marcoh at marcoh.net (Marco Hogewoning) Date: Fri, 27 Aug 2010 00:20:27 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <20100826191341.GB10681@srv03.cluenet.de> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <1282837386.31961.21.camel@daniel> <322E2633-C041-428F-92F3-25F38DAE6F5D@marcoh.net> <20100826191341.GB10681@srv03.cluenet.de> Message-ID: <8E2D0433-9E92-4623-AE34-66385DE19580@marcoh.net> On 26 aug 2010, at 21:13, Daniel Roesen wrote: > On Thu, Aug 26, 2010 at 06:52:37PM +0200, Marco Hogewoning wrote: >> Both, on one side we know we have a lot of non-IPv6 capable modems >> out there, > > Modems, or actually routers (with built-in modems)? Most of them are routers as well > I guess you do DHCP-PD for customer's LAN interface? What size of LAN > block does a customer get, just a /64? Or /56 and first /64 out of that > getting assigned to the LAN interface? PPP is unnumbered ie link local only, we assign a /48 per customer available as IA_PD. MarcoH From marcoh at marcoh.net Fri Aug 27 00:22:57 2010 From: marcoh at marcoh.net (Marco Hogewoning) Date: Fri, 27 Aug 2010 00:22:57 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <20100827072344.636bf9b5@opy.nosense.org> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <1282837386.31961.21.camel@daniel> <322E2633-C041-428F-92F3-25F38DAE6F5D@marcoh.net> <20100827072344.636bf9b5@opy.nosense.org> Message-ID: <9179FBF0-AB70-416D-896B-F563B214C68D@marcoh.net> On 26 aug 2010, at 23:53, Mark Smith wrote: > On Thu, 26 Aug 2010 18:52:37 +0200 > Marco Hogewoning wrote: > >> >> On 26 aug 2010, at 17:43, Daniel Verlouw wrote: >> >>> Excellent news, good job! >>> >>> On Thu, 2010-08-26 at 17:22 +0200, Marco Hogewoning wrote: >>>> XS4ALL customers can activate IPv6 on their account with a simple flick of a switch in the Service Centre on the company?s website. >>> >>> just out of curiosity why not turned on by default ? Router/modem >>> compatibility issues? A provisioning thing? Or afraid of helpdesk >>> overload with calls such as "my connection all of sudden doesn't work >>> anymore after you turned on this Eye Pee Vee Six thing" ? ;) >> >> >> Both, on one side we know we have a lot of non-IPv6 capable modems out there, so why waist resources on them. The other is indeed the fact taht if we push IPv6 to people without them being aware we might run into trouble. >> > > I've seen an IPv4 only modem lock up two minutes after establishing a > PPP session, and then having upstream LNS periodically try to kick off > IPV6CP again. I suppose a lock up is somewhat equivalent to a PPP > Protocol-Reject ;-) Our side is completely passive, so as long as the modem doesn't try to setup IPV6CP nothing will happen. But it still eats a routing entry, address assigment and the whole lot. More over at least now we can see somebody is using IPv6, makes support a whole lot easier. MarcoH From tore.anderson at redpill-linpro.com Fri Aug 27 07:36:06 2010 From: tore.anderson at redpill-linpro.com (Tore Anderson) Date: Fri, 27 Aug 2010 07:36:06 +0200 (CEST) Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> Message-ID: <1796409352.33006.1282887366119.JavaMail.root@claudius.linpro.no> * Marco Hogewoning > XS4ALL OFFERS IPv6 TO ALL CUSTOMERS Congratulations! I assume this is the world's first large rollout of universally available native IPv6 connectivity to residental end users? It does give you bragging rights to be sure! It would be interesting to see some statistics on how many users choose to enable it, and perhaps even more so how many of them disable it again. And which kind of IPv6-related problems (if any) are commonly reported to your help desk, and how they are handled. By the way, do you enable IPv6 by default for new customers? Many providers are struggling with figuring out how to roll out IPv6, it seems. So in addition to learning more about the customer response, it would be great to hear a more technical walk-through of how you did it, what problems you ran into, and how you solved them. May I suggest that you give a presentation at the upcoming RIPE meeting? Best regards, -- Tore Anderson Redpill Linpro AS - http://www.redpill-linpro.com/ Tel: +47 21 54 41 27 From steve at ipv6canada.com Fri Aug 27 07:53:21 2010 From: steve at ipv6canada.com (Steve Bertrand) Date: Fri, 27 Aug 2010 01:53:21 -0400 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> Message-ID: <4C7752D1.1000107@ipv6canada.com> On 2010.08.26 11:22, Marco Hogewoning wrote: > Sorry for the spam... Don't apologize... 'gratz!!! Steve From tedm at ipinc.net Fri Aug 27 08:21:58 2010 From: tedm at ipinc.net (Ted Mittelstaedt) Date: Thu, 26 Aug 2010 23:21:58 -0700 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <4C7752D1.1000107@ipv6canada.com> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> Message-ID: <4C775986.70307@ipinc.net> On 8/26/2010 10:53 PM, Steve Bertrand wrote: > On 2010.08.26 11:22, Marco Hogewoning wrote: >> Sorry for the spam... > > Don't apologize... > > 'gratz!!! > So, when we do the same thing in a few months is it OK for us to spam too? :-) Ted From steve at ipv6canada.com Fri Aug 27 08:41:04 2010 From: steve at ipv6canada.com (Steve Bertrand) Date: Fri, 27 Aug 2010 02:41:04 -0400 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <4C775986.70307@ipinc.net> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> <4C775986.70307@ipinc.net> Message-ID: <4C775E00.5090108@ipv6canada.com> On 2010.08.27 02:21, Ted Mittelstaedt wrote: > > > On 8/26/2010 10:53 PM, Steve Bertrand wrote: >> On 2010.08.26 11:22, Marco Hogewoning wrote: >>> Sorry for the spam... >> >> Don't apologize... >> >> 'gratz!!! >> > > So, when we do the same thing in a few months is > it OK for us to spam too? :-) yup, spam away and promote! I didn't see the message in my personal mailbox, so I accept. I have no problem accepting 'spam' regarding v6 deployments on a v6 list ;) Hell, I'd even be willing to accept into my personal inbox unrelated spam from *any* provider, given that they were ready and able to deploy v6 to *all* of their clients, particularly their resi's. Steve From roger.wiklund at gmail.com Fri Aug 27 08:48:53 2010 From: roger.wiklund at gmail.com (Roger Wiklund) Date: Fri, 27 Aug 2010 08:48:53 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <4C7752D1.1000107@ipv6canada.com> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> Message-ID: I know I'm comparing DSL to Ethernet, and your DSL service might work completely different, but I just wanted to see what your thoughts are. My building at home has fiber to the basement, connecting to a Cisco 3750 switch. Each apartment connects to that switch with Cat6. It appears that all residents that have internet connection is in the same VLAN, and in the same subnet. I assume my ISP has the standard edge switch security against arp spoofing/poisoning, dhcp spoofing etc. I don't know if they have private VLAN with proxy-arp or something similar, but lets say for the arguments sake that I can send traffic to my neighbor directly on layer 2 via the switch, as we are in the same subnet. "PPP is unnumbered ie link local only, we assign a /48 per customer available as IA_PD." My questions are: Do you mean that each individual customer gets a /48, and another customer in the same building gets another /48, I.E you have to route to be able to communicate between them. Instead of them being in the same /48 and potentially being able to communicate on layer 2. Why are you assigning a /48 instead of a /56 or /64. Is it just "because we can" or are there other benefits etc? Thanks! Regards Roger On Fri, Aug 27, 2010 at 7:53 AM, Steve Bertrand wrote: > On 2010.08.26 11:22, Marco Hogewoning wrote: >> Sorry for the spam... > > Don't apologize... > > 'gratz!!! > > Steve > From marcoh at marcoh.net Fri Aug 27 09:10:41 2010 From: marcoh at marcoh.net (Marco Hogewoning) Date: Fri, 27 Aug 2010 09:10:41 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <1796409352.33006.1282887366119.JavaMail.root@claudius.linpro.no> References: <1796409352.33006.1282887366119.JavaMail.root@claudius.linpro.no> Message-ID: On 27 aug 2010, at 07:36, Tore Anderson wrote: > * Marco Hogewoning > >> XS4ALL OFFERS IPv6 TO ALL CUSTOMERS > > Congratulations! I assume this is the world's first large rollout of > universally available native IPv6 connectivity to residental end users? Dunno, I haven't come across much large deployments. > It does give you bragging rights to be sure! > > It would be interesting to see some statistics on how many users choose > to enable it, and perhaps even more so how many of them disable it > again. And which kind of IPv6-related problems (if any) are commonly > reported to your help desk, and how they are handled. By the way, do > you enable IPv6 by default for new customers? Statistics are being collected, have to see how much we can share without getting issues with non-disclosures. We don't enable b default at the moment and there are 2 main reasons: - The modem FW is still in beta and would rather not ship it by default, turning it on without a CPE that runs v6 gives a lot of pollution in all the system and screw our stats. - We are pretty sure we won't see any problems, but better safe than sorry. If somebody chooses to enable v6 and runs into problems at least they know what changed. This is pretty new also for our support staff. Like you mentioned there ain't many companies running in this scale so we really don't know how much issues we will see, if any. The pilot gave some indication but those were techies with an above average knowledge level and direct access to engineering for support. Now it's Joe the plumber calling first line support and not even knowing what an IP address is :( > Many providers are struggling with figuring out how to roll out IPv6, > it seems. So in addition to learning more about the customer response, > it would be great to hear a more technical walk-through of how you did > it, what problems you ran into, and how you solved them. May I suggest > that you give a presentation at the upcoming RIPE meeting? It's on the list of things todo, I will consult with myself and my fellow chairs on an agenda slot :) MarcoH From marcoh at marcoh.net Fri Aug 27 09:16:09 2010 From: marcoh at marcoh.net (Marco Hogewoning) Date: Fri, 27 Aug 2010 09:16:09 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> Message-ID: On 27 aug 2010, at 08:48, Roger Wiklund wrote: > I know I'm comparing DSL to Ethernet, and your DSL service might work > completely different, but I just wanted to see what your thoughts > are. It isn't that different as we carry ethernet frames over DSL > My building at home has fiber to the basement, connecting to a Cisco > 3750 switch. Each apartment connects to that switch with Cat6. > It appears that all residents that have internet connection is in the > same VLAN, and in the same subnet. I assume my ISP has the standard > edge switch security against arp spoofing/poisoning, dhcp spoofing > etc. > > I don't know if they have private VLAN with proxy-arp or something > similar, but lets say for the arguments sake that I can send traffic > to my neighbor directly on layer 2 via the switch, as we are in the > same subnet. > > "PPP is unnumbered ie link local only, we assign a /48 per customer > available as IA_PD." Thas why we run PPPoE, unfortunately most equipment can't give the same type of protection for anti-spoofing, DHCP snooping and ND proxy for IPv6 as they do for IPv4 :( > My questions are: > > Do you mean that each individual customer gets a /48, and another > customer in the same building gets another /48, I.E you have to route > to be able to communicate between them. Instead of them being in the > same /48 and potentially being able to communicate on layer 2. Yups > Why are you assigning a /48 instead of a /56 or /64. Is it just > "because we can" or are there other benefits etc? Because we can and mostly because a one-size-fits all was much easier to implement at the moment. The whole address management system is automated, it is capable of supplying itself with new addresses when pools run out etc. Als a lot of folks out there once heard about the 48 and are still sticking to it, even when the policy differs. Saves ourself a lot of discussion and waving policy. MarcoH From dr at cluenet.de Fri Aug 27 09:49:24 2010 From: dr at cluenet.de (Daniel Roesen) Date: Fri, 27 Aug 2010 09:49:24 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> Message-ID: <20100827074924.GA4646@srv03.cluenet.de> On Fri, Aug 27, 2010 at 09:16:09AM +0200, Marco Hogewoning wrote: > Als a lot of folks out there once heard about the 48 and are still > sticking to it, even when the policy differs. Saves ourself a lot > of discussion and waving policy. Unless you run out of /48s and have to go back to RIPE NCC to get more address space. And then you're judged on /56s, not /48s for efficiency. Best regards, Daniel -- CLUE-RIPE -- Jabber: dr at cluenet.de -- dr at IRCnet -- PGP: 0xA85C8AA0 From marcoh at marcoh.net Fri Aug 27 09:55:28 2010 From: marcoh at marcoh.net (Marco Hogewoning) Date: Fri, 27 Aug 2010 09:55:28 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <20100827074924.GA4646@srv03.cluenet.de> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> <20100827074924.GA4646@srv03.cluenet.de> Message-ID: <41715C4A-E0DF-4E74-8184-DB8ADCDD0A01@marcoh.net> On 27 aug 2010, at 09:49, Daniel Roesen wrote: > On Fri, Aug 27, 2010 at 09:16:09AM +0200, Marco Hogewoning wrote: >> Als a lot of folks out there once heard about the 48 and are still >> sticking to it, even when the policy differs. Saves ourself a lot >> of discussion and waving policy. > > Unless you run out of /48s and have to go back to RIPE NCC to get more > address space. And then you're judged on /56s, not /48s for efficiency. No, you are judged on /56 equivalents....the /48 assignment is perfectly valid. MarcoH From jeroen at unfix.org Fri Aug 27 10:00:46 2010 From: jeroen at unfix.org (Jeroen Massar) Date: Fri, 27 Aug 2010 10:00:46 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> Message-ID: <4C7770AE.2020102@unfix.org> On 2010-08-26 17:22, Marco Hogewoning wrote: > Sorry for the spam... That is not spam ;) > Groet, > > MarcoH > > --- > > PRESS RELEASE 26 August 2010 > > IPv6 available for all, starting today Awesome. And I really like the way of having a knob that users can switch in the control panel*, that is the proper way to do it! * which they can also use for other things like turning on/off filtering netbios ports and a lot more cool stuff. Greets, Jeroen From shane at time-travellers.org Fri Aug 27 10:00:54 2010 From: shane at time-travellers.org (Shane Kerr) Date: Fri, 27 Aug 2010 10:00:54 +0200 Subject: /48 vs. /56 for customers, was XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <20100827074924.GA4646@srv03.cluenet.de> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> <20100827074924.GA4646@srv03.cluenet.de> Message-ID: <1282896054.14629.13934.camel@shane-asus-laptop> Daniel, On Fri, 2010-08-27 at 09:49 +0200, Daniel Roesen wrote: > On Fri, Aug 27, 2010 at 09:16:09AM +0200, Marco Hogewoning wrote: > > Als a lot of folks out there once heard about the 48 and are still > > sticking to it, even when the policy differs. Saves ourself a lot > > of discussion and waving policy. > > Unless you run out of /48s and have to go back to RIPE NCC to get more > address space. And then you're judged on /56s, not /48s for efficiency. Hm... that's a good point. With at "normal" /32 allocation you can only give 65k customers a /48, which is non-trivial but also almost certain to happen sooner or later. So any ISP of size will run out of address space if they use /48, and then will have to shift to /56 eventually. Best to use /56 initially? Perhaps this should be a strong recommendation somewhere? -- Shane From dr at cluenet.de Fri Aug 27 10:07:47 2010 From: dr at cluenet.de (Daniel Roesen) Date: Fri, 27 Aug 2010 10:07:47 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <41715C4A-E0DF-4E74-8184-DB8ADCDD0A01@marcoh.net> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> <20100827074924.GA4646@srv03.cluenet.de> <41715C4A-E0DF-4E74-8184-DB8ADCDD0A01@marcoh.net> Message-ID: <20100827080747.GC4646@srv03.cluenet.de> On Fri, Aug 27, 2010 at 09:55:28AM +0200, Marco Hogewoning wrote: > > Unless you run out of /48s and have to go back to RIPE NCC to get more > > address space. And then you're judged on /56s, not /48s for efficiency. > > No, you are judged on /56 equivalents....the /48 assignment is perfectly valid. Sure it is, but it's a difference in the amount of customers you can support, isn't it? With the old /48 HD ratio 0.80 policy, a /32 was to support a customer base of 52.429 before you qualified for additional address space. With the new /56 policy with HD ratio 0.94, RIPE NCC asks for 6.183.533 customers before the /32 allocation is deemed efficiently used. Is my understanding of the policy wrong? Best regards, Daniel -- CLUE-RIPE -- Jabber: dr at cluenet.de -- dr at IRCnet -- PGP: 0xA85C8AA0 From jeroen at unfix.org Fri Aug 27 10:14:45 2010 From: jeroen at unfix.org (Jeroen Massar) Date: Fri, 27 Aug 2010 10:14:45 +0200 Subject: /48 vs. /56 for customers, was XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <1282896054.14629.13934.camel@shane-asus-laptop> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> <20100827074924.GA4646@srv03.cluenet.de> <1282896054.14629.13934.camel@shane-asus-laptop> Message-ID: <4C7773F5.4030308@unfix.org> On 2010-08-27 10:00, Shane Kerr wrote: > Daniel, > > On Fri, 2010-08-27 at 09:49 +0200, Daniel Roesen wrote: >> On Fri, Aug 27, 2010 at 09:16:09AM +0200, Marco Hogewoning wrote: >>> Als a lot of folks out there once heard about the 48 and are still >>> sticking to it, even when the policy differs. Saves ourself a lot >>> of discussion and waving policy. >> >> Unless you run out of /48s and have to go back to RIPE NCC to get more >> address space. And then you're judged on /56s, not /48s for efficiency. > > Hm... that's a good point. > > With at "normal" /32 allocation you can only give 65k customers a /48, > which is non-trivial but also almost certain to happen sooner or later. > So any ISP of size will run out of address space if they use /48, and > then will have to shift to /56 eventually. > > Best to use /56 initially? Perhaps this should be a strong > recommendation somewhere? No, they just have to do a proper address plan when they request their allocation and get something bigger. I mean, it is not THAT hard, you have at the current moment X customers. You can maximally grow your market to Y customers if all the competition is gone, figure out the chance that that happens and you know what your largest set is ever going to be. Round the X up to how much you expect to be in say 5 years, and presto, convert that into number of /48's round that nicely to a factor of 2 and you have the number of bits that you need. Write a proper address plan and submit to the NCC. How else do you think two ISP's came up with /19's? Heck, if you are a government you can get a /13! ;) >From a conservation point of view though, if one would state "I give /56's to home-customers, /48's to companies" and that would be taken as a universal thing, I think it would be fine. Then again, does it really matter on the grand scale, 2000::/3 won't fill up that easily... (fast forward a 25 years, and that will bite me probably ;) Greets, Jeroen From dr at cluenet.de Fri Aug 27 10:20:19 2010 From: dr at cluenet.de (Daniel Roesen) Date: Fri, 27 Aug 2010 10:20:19 +0200 Subject: /48 vs. /56 for customers, was XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <4C7773F5.4030308@unfix.org> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> <20100827074924.GA4646@srv03.cluenet.de> <1282896054.14629.13934.camel@shane-asus-laptop> <4C7773F5.4030308@unfix.org> Message-ID: <20100827082019.GD4646@srv03.cluenet.de> On Fri, Aug 27, 2010 at 10:14:45AM +0200, Jeroen Massar wrote: > How else do you think two ISP's came up with /19's? That was in times where the /48 HD 0.8 policy was in effect. Best regards, Daniel (designer of the addressing plan of a /21 allocation) -- CLUE-RIPE -- Jabber: dr at cluenet.de -- dr at IRCnet -- PGP: 0xA85C8AA0 From gert at space.net Fri Aug 27 10:20:56 2010 From: gert at space.net (Gert Doering) Date: Fri, 27 Aug 2010 10:20:56 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> Message-ID: <20100827082056.GE61734@Space.Net> hi, On Fri, Aug 27, 2010 at 08:48:53AM +0200, Roger Wiklund wrote: > Do you mean that each individual customer gets a /48, and another > customer in the same building gets another /48, I.E you have to route > to be able to communicate between them. Instead of them being in the > same /48 and potentially being able to communicate on layer 2. That would make sense. After all, you do not want your inside network to be in the same network as your neighbour's hosts, do you? Gert Doering -- NetMaster -- Total number of prefixes smaller than registry allocations: 155817 SpaceNet AG Vorstand: Sebastian v. Bomhard Joseph-Dollinger-Bogen 14 Aufsichtsratsvors.: A. Grundner-Culemann D-80807 Muenchen HRB: 136055 (AG Muenchen) Tel: +49 (89) 32356-444 USt-IdNr.: DE813185279 From gert at space.net Fri Aug 27 10:21:33 2010 From: gert at space.net (Gert Doering) Date: Fri, 27 Aug 2010 10:21:33 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <20100827074924.GA4646@srv03.cluenet.de> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> <20100827074924.GA4646@srv03.cluenet.de> Message-ID: <20100827082133.GF61734@Space.Net> Hi, On Fri, Aug 27, 2010 at 09:49:24AM +0200, Daniel Roesen wrote: > On Fri, Aug 27, 2010 at 09:16:09AM +0200, Marco Hogewoning wrote: > > Als a lot of folks out there once heard about the 48 and are still > > sticking to it, even when the policy differs. Saves ourself a lot > > of discussion and waving policy. > > Unless you run out of /48s and have to go back to RIPE NCC to get more > address space. And then you're judged on /56s, not /48s for efficiency. An /48 assigned counts as "all the /56s in there are assigned". Gert Doering -- NetMaster -- Total number of prefixes smaller than registry allocations: 155817 SpaceNet AG Vorstand: Sebastian v. Bomhard Joseph-Dollinger-Bogen 14 Aufsichtsratsvors.: A. Grundner-Culemann D-80807 Muenchen HRB: 136055 (AG Muenchen) Tel: +49 (89) 32356-444 USt-IdNr.: DE813185279 From dr at cluenet.de Fri Aug 27 10:38:19 2010 From: dr at cluenet.de (Daniel Roesen) Date: Fri, 27 Aug 2010 10:38:19 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <20100827082133.GF61734@Space.Net> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> <20100827074924.GA4646@srv03.cluenet.de> <20100827082133.GF61734@Space.Net> Message-ID: <20100827083819.GE4646@srv03.cluenet.de> On Fri, Aug 27, 2010 at 10:21:33AM +0200, Gert Doering wrote: > > Unless you run out of /48s and have to go back to RIPE NCC to get more > > address space. And then you're judged on /56s, not /48s for efficiency. > > An /48 assigned counts as "all the /56s in there are assigned". Yes, that's the problem. For just one customer. Best regards, Daniel -- CLUE-RIPE -- Jabber: dr at cluenet.de -- dr at IRCnet -- PGP: 0xA85C8AA0 From nick at foobar.org Fri Aug 27 10:51:35 2010 From: nick at foobar.org (Nick Hilliard) Date: Fri, 27 Aug 2010 09:51:35 +0100 Subject: /48 vs. /56 for customers, was XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <4C7773F5.4030308@unfix.org> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> <20100827074924.GA4646@srv03.cluenet.de> <1282896054.14629.13934.camel@shane-asus-laptop> <4C7773F5.4030308@unfix.org> Message-ID: <4C777C97.6070500@foobar.org> On 27/08/2010 09:14, Jeroen Massar wrote: > Heck, if you are a government you can get a /13! ;) We all know that this statement is not accurate :-) Nick From gert at space.net Fri Aug 27 10:52:31 2010 From: gert at space.net (Gert Doering) Date: Fri, 27 Aug 2010 10:52:31 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <20100827083819.GE4646@srv03.cluenet.de> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> <20100827074924.GA4646@srv03.cluenet.de> <20100827082133.GF61734@Space.Net> <20100827083819.GE4646@srv03.cluenet.de> Message-ID: <20100827085231.GG61734@Space.Net> Hi, On Fri, Aug 27, 2010 at 10:38:19AM +0200, Daniel Roesen wrote: > On Fri, Aug 27, 2010 at 10:21:33AM +0200, Gert Doering wrote: > > > Unless you run out of /48s and have to go back to RIPE NCC to get more > > > address space. And then you're judged on /56s, not /48s for efficiency. > > > > An /48 assigned counts as "all the /56s in there are assigned". > > Yes, that's the problem. For just one customer. What's the *problem* here? If you hand out your /32 in chunks of /48s to your customers, and the /32 is full, it is *full* - and you're entitled to get more space. So why is this considered a problem? (For us, we're likely never going to fill our /32, so it doesn't really matter what size of block we hand out to customers - and for networks where this matters, it really should be a matter of having a proper address plan presented to the RIPE NCC. Now, if the NCC denies a request based on "we don't do /48s here, assign /56s to your customers!" that would certainly something to escalate) Gert Doering -- NetMaster -- Total number of prefixes smaller than registry allocations: 155817 SpaceNet AG Vorstand: Sebastian v. Bomhard Joseph-Dollinger-Bogen 14 Aufsichtsratsvors.: A. Grundner-Culemann D-80807 Muenchen HRB: 136055 (AG Muenchen) Tel: +49 (89) 32356-444 USt-IdNr.: DE813185279 From jeroen at unfix.org Fri Aug 27 11:02:20 2010 From: jeroen at unfix.org (Jeroen Massar) Date: Fri, 27 Aug 2010 11:02:20 +0200 Subject: /48 vs. /56 for customers, was XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <4C777C97.6070500@foobar.org> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> <20100827074924.GA4646@srv03.cluenet.de> <1282896054.14629.13934.camel@shane-asus-laptop> <4C7773F5.4030308@unfix.org> <4C777C97.6070500@foobar.org> Message-ID: <4C777F1C.7050303@unfix.org> On 2010-08-27 10:51, Nick Hilliard wrote: > On 27/08/2010 09:14, Jeroen Massar wrote: >> Heck, if you are a government you can get a /13! ;) > > We all know that this statement is not accurate :-) True true, there are several other /32's that are also allocated to that same government. Greets, Jeroen From roger.wiklund at gmail.com Fri Aug 27 11:06:34 2010 From: roger.wiklund at gmail.com (Roger Wiklund) Date: Fri, 27 Aug 2010 11:06:34 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <20100827082056.GE61734@Space.Net> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> <20100827082056.GE61734@Space.Net> Message-ID: Indeed. It's just that for many IPv4 deployments, users in the same building/part of town usually gets an IP in the same subnet. But that of course is based on today's networks where you have a private network in your LAN and NAT etc. It's like how its recommended to use /64 on point-to-point links, that's just crazy waste of addresses to me. And to assign as single user a /48, I mean that user will never ever use up all the IPs. But I guess I need to start thinking IPv6 :) On Fri, Aug 27, 2010 at 10:20 AM, Gert Doering wrote: > hi, > > On Fri, Aug 27, 2010 at 08:48:53AM +0200, Roger Wiklund wrote: >> Do you mean that each individual customer gets a /48, and another >> customer in the same building gets another /48, I.E you have to route >> to be able to communicate between them. Instead of them being in the >> same /48 and potentially being able to communicate on layer 2. > > That would make sense. > > After all, you do not want your inside network to be in the same network > as your neighbour's hosts, do you? > > Gert Doering > ? ? ? ?-- NetMaster > -- > Total number of prefixes smaller than registry allocations: ?155817 > > SpaceNet AG ? ? ? ? ? ? ? ? ? ? ? ?Vorstand: Sebastian v. Bomhard > Joseph-Dollinger-Bogen 14 ? ? ? ? ?Aufsichtsratsvors.: A. Grundner-Culemann > D-80807 Muenchen ? ? ? ? ? ? ? ? ? HRB: 136055 (AG Muenchen) > Tel: +49 (89) 32356-444 ? ? ? ? ? ?USt-IdNr.: DE813185279 > From marcoh at marcoh.net Fri Aug 27 11:47:27 2010 From: marcoh at marcoh.net (Marco Hogewoning) Date: Fri, 27 Aug 2010 11:47:27 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <20100827082133.GF61734@Space.Net> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> <20100827074924.GA4646@srv03.cluenet.de> <20100827082133.GF61734@Space.Net> Message-ID: <01148115-69AC-417C-A397-C1F13194AEBB@marcoh.net> On 27 aug 2010, at 10:21, Gert Doering wrote: > Hi, > > On Fri, Aug 27, 2010 at 09:49:24AM +0200, Daniel Roesen wrote: >> On Fri, Aug 27, 2010 at 09:16:09AM +0200, Marco Hogewoning wrote: >>> Als a lot of folks out there once heard about the 48 and are still >>> sticking to it, even when the policy differs. Saves ourself a lot >>> of discussion and waving policy. >> >> Unless you run out of /48s and have to go back to RIPE NCC to get more >> address space. And then you're judged on /56s, not /48s for efficiency. > > An /48 assigned counts as "all the /56s in there are assigned". > That's the way I understood it as welland so far communications with IPRA indicate this is also the way it's handled. The bigger problem I have is how to handle sec 5.5 of RIPE-481. But I have a trusted source who tells this is being worked on. Groet, MarcoH From gert at space.net Fri Aug 27 11:49:53 2010 From: gert at space.net (Gert Doering) Date: Fri, 27 Aug 2010 11:49:53 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> <20100827082056.GE61734@Space.Net> Message-ID: <20100827094953.GJ61734@Space.Net> Hi, On Fri, Aug 27, 2010 at 11:06:34AM +0200, Roger Wiklund wrote: > It's like how its recommended to use /64 on point-to-point links, > that's just crazy waste of addresses to me. And to assign as single > user a /48, I mean that user will never ever use up all the IPs. You could assign the user a /96, and he would never be able to use up all the addresses in there either :-) > But I guess I need to start thinking IPv6 :) Indeed - just get rid of your IPv4 mindset :-) (Unfortunately, that takes a lot of unlearning) Gert Doering -- NetMaster -- Total number of prefixes smaller than registry allocations: 155817 SpaceNet AG Vorstand: Sebastian v. Bomhard Joseph-Dollinger-Bogen 14 Aufsichtsratsvors.: A. Grundner-Culemann D-80807 Muenchen HRB: 136055 (AG Muenchen) Tel: +49 (89) 32356-444 USt-IdNr.: DE813185279 -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 306 bytes Desc: not available Url : http://lists.cluenet.de/pipermail/ipv6-ops/attachments/20100827/88264093/attachment.bin From gert at space.net Fri Aug 27 11:57:07 2010 From: gert at space.net (Gert Doering) Date: Fri, 27 Aug 2010 11:57:07 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <01148115-69AC-417C-A397-C1F13194AEBB@marcoh.net> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> <20100827074924.GA4646@srv03.cluenet.de> <20100827082133.GF61734@Space.Net> <01148115-69AC-417C-A397-C1F13194AEBB@marcoh.net> Message-ID: <20100827095707.GK61734@Space.Net> Hi, On Fri, Aug 27, 2010 at 11:47:27AM +0200, Marco Hogewoning wrote: > The bigger problem I have is how to handle sec 5.5 of RIPE-481. But > I have a trusted source who tells this is being worked on. Another trusted source tells me that indeed, the language of that section needs major cleanup - and part of that will be to understand what we *want* to be there. (Which makes this an interesting cross-WG excercise). My current interpretation of this is "don't register your end user data unless they ask for it, but be prepared to send a huge database dump to the IPRAs when they come asking". Which seems to be somewhat in line with the APWG and DBWG interpretation of things :-) - but the language of this document is very old, from the "initial global policy", and has never been adapted to RIPE terminology. (My trusted source asks me to tell your trusted source to not send word documents to mailing lists if "plain ascii" is all that's in there :-) ) Gert Doering -- NetMaster -- Total number of prefixes smaller than registry allocations: 155817 SpaceNet AG Vorstand: Sebastian v. Bomhard Joseph-Dollinger-Bogen 14 Aufsichtsratsvors.: A. Grundner-Culemann D-80807 Muenchen HRB: 136055 (AG Muenchen) Tel: +49 (89) 32356-444 USt-IdNr.: DE813185279 -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 306 bytes Desc: not available Url : http://lists.cluenet.de/pipermail/ipv6-ops/attachments/20100827/6e830882/attachment.bin From bjorn at mork.no Fri Aug 27 14:19:43 2010 From: bjorn at mork.no (=?utf-8?Q?Bj=C3=B8rn_Mork?=) Date: Fri, 27 Aug 2010 14:19:43 +0200 Subject: /48 vs. /56 for customers, was XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <1282896054.14629.13934.camel@shane-asus-laptop> (Shane Kerr's message of "Fri, 27 Aug 2010 10:00:54 +0200") References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> <20100827074924.GA4646@srv03.cluenet.de> <1282896054.14629.13934.camel@shane-asus-laptop> Message-ID: <87fwy06ye8.fsf@nemi.mork.no> Shane Kerr writes: > With at "normal" /32 allocation you can only give 65k customers a /48, > which is non-trivial but also almost certain to happen sooner or later. > So any ISP of size will run out of address space if they use /48, and > then will have to shift to /56 eventually. Any ISP of size will request and get much more than a /32. Bj?rn From bjorn at mork.no Fri Aug 27 14:26:38 2010 From: bjorn at mork.no (=?utf-8?Q?Bj=C3=B8rn_Mork?=) Date: Fri, 27 Aug 2010 14:26:38 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <8E2D0433-9E92-4623-AE34-66385DE19580@marcoh.net> (Marco Hogewoning's message of "Fri, 27 Aug 2010 00:20:27 +0200") References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <1282837386.31961.21.camel@daniel> <322E2633-C041-428F-92F3-25F38DAE6F5D@marcoh.net> <20100826191341.GB10681@srv03.cluenet.de> <8E2D0433-9E92-4623-AE34-66385DE19580@marcoh.net> Message-ID: <87bp8o6y2p.fsf@nemi.mork.no> Marco Hogewoning writes: > On 26 aug 2010, at 21:13, Daniel Roesen wrote: > >> On Thu, Aug 26, 2010 at 06:52:37PM +0200, Marco Hogewoning wrote: >>> Both, on one side we know we have a lot of non-IPv6 capable modems >>> out there, >> >> Modems, or actually routers (with built-in modems)? > > Most of them are routers as well > >> I guess you do DHCP-PD for customer's LAN interface? What size of LAN >> block does a customer get, just a /64? Or /56 and first /64 out of that >> getting assigned to the LAN interface? > > > PPP is unnumbered ie link local only, Why? How do you do CPE mangament? Steal a part of the delegated prefix? Or using IPv4 only? Or not at all? > we assign a /48 per customer available as IA_PD. Great! Thanks for pushing this. It will make it easier for the rest of us to just follow. Do you know which DHCPv6 client the Fritzbox uses? Anything open source, or their own implementation? Bj?rn From marcoh at marcoh.net Fri Aug 27 14:30:13 2010 From: marcoh at marcoh.net (Marco Hogewoning) Date: Fri, 27 Aug 2010 14:30:13 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <87bp8o6y2p.fsf@nemi.mork.no> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <1282837386.31961.21.camel@daniel> <322E2633-C041-428F-92F3-25F38DAE6F5D@marcoh.net> <20100826191341.GB10681@srv03.cluenet.de> <8E2D0433-9E92-4623-AE34-66385DE19580@marcoh.net> <87bp8o6y2p.fsf@nemi.mork.no> Message-ID: On 27 aug 2010, at 14:26, Bj?rn Mork wrote: > Marco Hogewoning writes: >> On 26 aug 2010, at 21:13, Daniel Roesen wrote: >> >>> On Thu, Aug 26, 2010 at 06:52:37PM +0200, Marco Hogewoning wrote: >>>> Both, on one side we know we have a lot of non-IPv6 capable modems >>>> out there, >>> >>> Modems, or actually routers (with built-in modems)? >> >> Most of them are routers as well >> >>> I guess you do DHCP-PD for customer's LAN interface? What size of LAN >>> block does a customer get, just a /64? Or /56 and first /64 out of that >>> getting assigned to the LAN interface? >> >> >> PPP is unnumbered ie link local only, > > Why? How do you do CPE mangament? Steal a part of the delegated > prefix? Or using IPv4 only? Or not at all? Not at all, CPE is controlled by the customer. >> we assign a /48 per customer available as IA_PD. > > Great! Thanks for pushing this. It will make it easier for the rest of > us to just follow. > > Do you know which DHCPv6 client the Fritzbox uses? Anything open > source, or their own implementation? Don't know, probably some own sttuff piggy backed on an open source base. Groet, MarcoH From swmike at swm.pp.se Fri Aug 27 15:56:17 2010 From: swmike at swm.pp.se (Mikael Abrahamsson) Date: Fri, 27 Aug 2010 15:56:17 +0200 (CEST) Subject: /48 vs. /56 for customers, was XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <87fwy06ye8.fsf@nemi.mork.no> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> <20100827074924.GA4646@srv03.cluenet.de> <1282896054.14629.13934.camel@shane-asus-laptop> <87fwy06ye8.fsf@nemi.mork.no> Message-ID: On Fri, 27 Aug 2010, Bj?rn Mork wrote: > Any ISP of size will request and get much more than a /32. Yes, exactly. I don't understand why people obsess abot the initial default /32. We have a /25, we returned our initial (from 2001 or so) /32, so we could make a proper addressing plan. I don't understand why others aren't doing the same. -- Mikael Abrahamsson email: swmike at swm.pp.se From nick at foobar.org Fri Aug 27 16:02:48 2010 From: nick at foobar.org (Nick Hilliard) Date: Fri, 27 Aug 2010 16:02:48 +0200 Subject: /48 vs. /56 for customers, was XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <4C777F1C.7050303@unfix.org> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> <20100827074924.GA4646@srv03.cluenet.de> <1282896054.14629.13934.camel@shane-asus-laptop> <4C7773F5.4030308@unfix.org> <4C777C97.6070500@foobar.org> <4C777F1C.7050303@unfix.org> Message-ID: <4C77C588.2090002@foobar.org> On 27/08/2010 11:02, Jeroen Massar wrote: > True true, there are several other /32's that are also allocated to that > same government. I'm sure there are plenty. The point is that the DoD were never assigned or allocated a /13 by ARIN or anyone else. To suggest otherwise is factually incorrect. Nick From jeroen at unfix.org Fri Aug 27 16:15:41 2010 From: jeroen at unfix.org (Jeroen Massar) Date: Fri, 27 Aug 2010 16:15:41 +0200 Subject: /48 vs. /56 for customers, was XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <4C77C588.2090002@foobar.org> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> <20100827074924.GA4646@srv03.cluenet.de> <1282896054.14629.13934.camel@shane-asus-laptop> <4C7773F5.4030308@unfix.org> <4C777C97.6070500@foobar.org> <4C777F1C.7050303@unfix.org> <4C77C588.2090002@foobar.org> Message-ID: <4C77C88D.1030007@unfix.org> On 2010-08-27 16:02, Nick Hilliard wrote: > On 27/08/2010 11:02, Jeroen Massar wrote: >> True true, there are several other /32's that are also allocated to that >> same government. > > I'm sure there are plenty. > > The point is that the DoD were never assigned or allocated a /13 by ARIN > or anyone else. To suggest otherwise is factually incorrect. It is not incorrect as the space in between their /22's will never ever be used by anybody else especially as it most very likely is reserved for them anyway. Greets, Jeroen From nick at foobar.org Fri Aug 27 16:32:55 2010 From: nick at foobar.org (Nick Hilliard) Date: Fri, 27 Aug 2010 15:32:55 +0100 Subject: /48 vs. /56 for customers, was XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <4C77C88D.1030007@unfix.org> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> <20100827074924.GA4646@srv03.cluenet.de> <1282896054.14629.13934.camel@shane-asus-laptop> <4C7773F5.4030308@unfix.org> <4C777C97.6070500@foobar.org> <4C777F1C.7050303@unfix.org> <4C77C588.2090002@foobar.org> <4C77C88D.1030007@unfix.org> Message-ID: <4C77CC97.6000005@foobar.org> On 27/08/2010 15:15, Jeroen Massar wrote: > It is not incorrect as the space in between their /22's will never ever > be used by anybody else You have no evidence of this. > especially as it most very likely is reserved > for them anyway. You have no evidence of this either. I understand that you have a strong opinion on whether or not the DoD was assigned a /13. However, it is an opinion and not a fact. Nick From leo.vegoda at icann.org Fri Aug 27 18:41:09 2010 From: leo.vegoda at icann.org (Leo Vegoda) Date: Fri, 27 Aug 2010 09:41:09 -0700 Subject: /48 vs. /56 for customers, was XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> <20100827074924.GA4646@srv03.cluenet.de> <1282896054.14629.13934.camel@shane-asus-laptop> <87fwy06ye8.fsf@nemi.mork.no> Message-ID: On 27 Aug 2010, at 6:56, Mikael Abrahamsson wrote: > On Fri, 27 Aug 2010, Bj?rn Mork wrote: > >> Any ISP of size will request and get much more than a /32. > > Yes, exactly. I don't understand why people obsess abot the initial > default /32. > > We have a /25, we returned our initial (from 2001 or so) /32, so we could > make a proper addressing plan. I don't understand why others aren't doing > the same. Years ago, Arno Meulenkamp did an interesting analysis of what amount of small broadband ISPs overall might need for their IPv6 address space requirements (http://www.ripe.net/ripe/meetings/ripe-47/presentations/ripe47-ap-ipv6-needs.pdf). It would be interesting to review that work and do a new analysis based on more recent experience and early deployments. Regards, Leo From spz at serpens.de Fri Aug 27 21:09:40 2010 From: spz at serpens.de (S.P.Zeidler) Date: Fri, 27 Aug 2010 21:09:40 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <4C775986.70307@ipinc.net> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> <4C775986.70307@ipinc.net> Message-ID: <20100827190939.GB26262@serpens.de> Hi, Thus wrote Ted Mittelstaedt (tedm at ipinc.net): > So, when we do the same thing in a few months is > it OK for us to spam too? :-) The later you are the shorter your message must be :> regards, spz -- spz at serpens.de (S.P.Zeidler) From jeroen at unfix.org Sat Aug 28 00:33:18 2010 From: jeroen at unfix.org (Jeroen Massar) Date: Sat, 28 Aug 2010 00:33:18 +0200 Subject: /48 vs. /56 for customers, was XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <4C77CC97.6000005@foobar.org> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> <20100827074924.GA4646@srv03.cluenet.de> <1282896054.14629.13934.camel@shane-asus-laptop> <4C7773F5.4030308@unfix.org> <4C777C97.6070500@foobar.org> <4C777F1C.7050303@unfix.org> <4C77C588.2090002@foobar.org> <4C77C88D.1030007@unfix.org> <4C77CC97.6000005@foobar.org> Message-ID: <4C783D2E.8000300@unfix.org> On 2010-08-27 16:32, Nick Hilliard wrote: > On 27/08/2010 15:15, Jeroen Massar wrote: >> It is not incorrect as the space in between their /22's will never ever >> be used by anybody else > > You have no evidence of this. I am no cop or judge thus it might not be evidence but it is obvious that ARIN reserves blocks in between those /22 for potential growth next to the fact that none of the spaces in between have been allocated yet. Another nice question is of course also why ARIN didn't even bother to aggregate this into one huge chunk either. Then again RIPE also sometimes allocates 2x /32 instead of a single /31 and it happens also at other RIRs. Fun problem with that of course is that the RIR states already that these prefixes are not one huge chunk but should just be announced separately into BGP while if they would allocate the aggregate they could just be done with a single prefix in BGP. >> especially as it most very likely is reserved >> for them anyway. > > You have no evidence of this either. > > I understand that you have a strong opinion on whether or not the DoD > was assigned a /13. However, it is an opinion and not a fact. It seems you have a rather strong urge to try and cover it too, I wonder why. If you would not have such a strong opinion you would just state that you and me disagree and that is it. That is a fact ;) Greets, Jeroen (anyone some popcorn?.... ;) From dougb at dougbarton.us Sat Aug 28 00:53:40 2010 From: dougb at dougbarton.us (Doug Barton) Date: Fri, 27 Aug 2010 15:53:40 -0700 Subject: /48 vs. /56 for customers, was XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <4C783D2E.8000300@unfix.org> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> <20100827074924.GA4646@srv03.cluenet.de> <1282896054.14629.13934.camel@shane-asus-laptop> <4C7773F5.4030308@unfix.org> <4C777C97.6070500@foobar.org> <4C777F1C.7050303@unfix.org> <4C77C588.2090002@foobar.org> <4C77C88D.1030007@unfix.org> <4C77CC97.6000005@foobar.org> <4C783D2E.8000300@unfix.org> Message-ID: <4C7841F4.9020900@dougbarton.us> On 08/27/2010 03:33 PM, Jeroen Massar wrote: > you and me disagree and that is it How do I configure my routers for this? -- Improve the effectiveness of your Internet presence with a domain name makeover! http://SupersetSolutions.com/ Computers are useless. They can only give you answers. -- Pablo Picasso From dr at cluenet.de Sun Aug 29 10:19:58 2010 From: dr at cluenet.de (Daniel Roesen) Date: Sun, 29 Aug 2010 10:19:58 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <20100827085231.GG61734@Space.Net> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> <20100827074924.GA4646@srv03.cluenet.de> <20100827082133.GF61734@Space.Net> <20100827083819.GE4646@srv03.cluenet.de> <20100827085231.GG61734@Space.Net> Message-ID: <20100829081958.GA28141@srv03.cluenet.de> On Fri, Aug 27, 2010 at 10:52:31AM +0200, Gert Doering wrote: > On Fri, Aug 27, 2010 at 10:38:19AM +0200, Daniel Roesen wrote: > > On Fri, Aug 27, 2010 at 10:21:33AM +0200, Gert Doering wrote: > > > > Unless you run out of /48s and have to go back to RIPE NCC to get more > > > > address space. And then you're judged on /56s, not /48s for efficiency. > > > > > > An /48 assigned counts as "all the /56s in there are assigned". > > > > Yes, that's the problem. For just one customer. > > What's the *problem* here? If you hand out your /32 in chunks of /48s > to your customers, and the /32 is full, it is *full* - and you're entitled > to get more space. I don't see this in the policy. What's the deal with /56s then in the policy? > So why is this considered a problem? Please review the efficiency requirement using HD ratio. This requirement effectively sets a minimum amount of customers supported by a /32. ============================================================== 5.2.1. Subsequent allocation criteria Subsequent allocation will be provided when an organisation (i.e. ISP/LIR) satisfies the evaluation threshold of past address utilisation in terms of the number of sites in units of /56 assignments. The HD-Ratio [RFC 3194] is used to determine the utilisation thresholds that justify the allocation of additional address as described below. 5.2.2. Applied HD-Ratio The HD-Ratio value of 0.94 is adopted as indicating an acceptable address utilisation for justifying the allocation of additional address space. Appendix A provides a table showing the number of assignments that are necessary to achieve an acceptable utilisation value for a given address block size. ============================================================== ..."the NUMBER OF ASSIGNMENTS"... It's a difference wether I'm eligliable for more having made 2^16 or 2^32 assignments. Using /48 blocks, I will never reach the required 6183533 assignments because 6183533 > 2^16. I see no specification in the policy on how many /48 assignments are required to be deemed qualified for a new/extended allocation. Same goes for any other size between /56 and /48. Best regards, Daniel -- CLUE-RIPE -- Jabber: dr at cluenet.de -- dr at IRCnet -- PGP: 0xA85C8AA0 From meulenkamp at isoc.org Mon Aug 30 11:25:00 2010 From: meulenkamp at isoc.org (Arno Meulenkamp) Date: Mon, 30 Aug 2010 05:25:00 -0400 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <20100829081958.GA28141@srv03.cluenet.de> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> <20100827074924.GA4646@srv03.cluenet.de> <20100827082133.GF61734@Space.Net> <20100827083819.GE4646@srv03.cluenet.de> <20100827085231.GG61734@Space.Net> <20100829081958.GA28141@srv03.cluenet.de> Message-ID: <865E26E1-323D-4019-B236-9738CDC921B4@isoc.org> On 29 Aug 2010, at 04:19 , Daniel Roesen wrote: > On Fri, Aug 27, 2010 at 10:52:31AM +0200, Gert Doering wrote: >> On Fri, Aug 27, 2010 at 10:38:19AM +0200, Daniel Roesen wrote: >>> On Fri, Aug 27, 2010 at 10:21:33AM +0200, Gert Doering wrote: >>>>> Unless you run out of /48s and have to go back to RIPE NCC to get more >>>>> address space. And then you're judged on /56s, not /48s for efficiency. >>>> >>>> An /48 assigned counts as "all the /56s in there are assigned". >>> >>> Yes, that's the problem. For just one customer. >> >> What's the *problem* here? If you hand out your /32 in chunks of /48s >> to your customers, and the /32 is full, it is *full* - and you're entitled >> to get more space. > > I don't see this in the policy. What's the deal with /56s then in the > policy? They simply use /56 for the calculation of the HD ratio. If you assign a /48, effectively, you've assigned 256 /56 blocks, as counting for the HD ratio. >> So why is this considered a problem? > Please review the efficiency requirement using HD ratio. This > requirement effectively sets a minimum amount of customers supported by > a /32. > > ============================================================== > 5.2.1. Subsequent allocation criteria > > Subsequent allocation will be provided when an organisation (i.e. > ISP/LIR) satisfies the evaluation threshold of past address utilisation > in terms of the number of sites in units of /56 assignments. The > HD-Ratio [RFC 3194] is used to determine the utilisation thresholds that > justify the allocation of additional address as described below. It might be good to rephrase the text to "Subsequent allocation will be provided when an organisation (i.e. ISP/LIR) satisfies the evaluation threshold of past address utilisation in terms of assigned address space in units of /56 blocks." That might bypass this confusion. > I see no specification in the policy on how many /48 assignments are > required to be deemed qualified for a new/extended allocation. Same goes > for any other size between /56 and /48. ciao, Arno From gert at space.net Tue Aug 31 20:29:28 2010 From: gert at space.net (Gert Doering) Date: Tue, 31 Aug 2010 20:29:28 +0200 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <865E26E1-323D-4019-B236-9738CDC921B4@isoc.org> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> <4C7752D1.1000107@ipv6canada.com> <20100827074924.GA4646@srv03.cluenet.de> <20100827082133.GF61734@Space.Net> <20100827083819.GE4646@srv03.cluenet.de> <20100827085231.GG61734@Space.Net> <20100829081958.GA28141@srv03.cluenet.de> <865E26E1-323D-4019-B236-9738CDC921B4@isoc.org> Message-ID: <20100831182928.GE61734@Space.Net> Hi Arno, thanks for answering this - I planned to dig into the documents to see what the exact wording was but was away for two days. On Mon, Aug 30, 2010 at 05:25:00AM -0400, Arno Meulenkamp wrote: [..] > > I don't see this in the policy. What's the deal with /56s then in the > > policy? > > They simply use /56 for the calculation of the HD ratio. If you assign > a /48, effectively, you've assigned 256 /56 blocks, as counting for > the HD ratio. Very clear :-) > > ============================================================== > > 5.2.1. Subsequent allocation criteria > > > > Subsequent allocation will be provided when an organisation (i.e. > > ISP/LIR) satisfies the evaluation threshold of past address utilisation > > in terms of the number of sites in units of /56 assignments. The > > HD-Ratio [RFC 3194] is used to determine the utilisation thresholds that > > justify the allocation of additional address as described below. > > It might be good to rephrase the text to "Subsequent allocation will > be provided when an organisation (i.e. ISP/LIR) satisfies the evaluation > threshold of past address utilisation in terms of assigned address > space in units of /56 blocks." > > That might bypass this confusion. Yes, this sounds good. I take this from there to APWG to discuss and get the text clarified (most likely, run the change through the formal PD process, as it *is* a change to the policy documents...) Gert Doering -- APWG chair -- Total number of prefixes smaller than registry allocations: 155817 SpaceNet AG Vorstand: Sebastian v. Bomhard Joseph-Dollinger-Bogen 14 Aufsichtsratsvors.: A. Grundner-Culemann D-80807 Muenchen HRB: 136055 (AG Muenchen) Tel: +49 (89) 32356-444 USt-IdNr.: DE813185279 From Niall.oReilly at ucd.ie Fri Aug 27 12:27:52 2010 From: Niall.oReilly at ucd.ie (Niall O'Reilly) Date: Fri, 27 Aug 2010 11:27:52 +0100 Subject: XS4ALL Introduces native IPv6 for DSL customers In-Reply-To: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> References: <70B48AD4-2913-447F-8AC3-617E03779F38@marcoh.net> Message-ID: <4C779328.1050106@ucd.ie> On 26/08/10 16:22, Marco Hogewoning wrote: > IPv6 available for all, starting today More congrats! /Niall