Please find below all previous versions of the IPv6 CPE survey and matrix with the most recent version on top.
The 2012 IPv6 CPE Survey (September 2012)
IPv6 Survey Results - May 2011
IPv6 CPE Survey - Please Participate (April 2011)
The Future of the IPv6 CPE Survey - Your Input Needed (January 2011)
IPv6 CPE Survey - Updated (January 2011)
IPv6 CPE Survey - Updated (September 2010)
IPv6 CPE Survey - Updated (July 2010)
Comments are disabled on articles published more than a year ago. If you'd like to inform us of any issues, please reach out to us via the contact form here.
ipv6 not "Labor only" or beta anymore <br /><br />Fritzboxes 7390, 7320, 7270(v2, v3), 7240, 6360 Cable, 3270 and may be more<br />support IPv6 in recent released firmware
Hide 2 replies
Yes, we are aware of that. It will be included in the next iteration of the matrix. But more importantly we are curious about experiences with this new code, especially on the cable side of things.<br /><br />Thanks
Hide one reply
Here's some IPv6 issues I've found with the AVM code for the 7390 -<br /><br />o IPv6 full rate throughput half IPv4's at 100Mbps in the upstream <br />direction over an EWAN / PPPoE service.<br /><br />o Curr Hop Count in RAs set to 255, not 64 or zero<br /><br />o MTU of 1460 announced in RAs, not PPPoE MTU, or left as LAN default so PMTU takes care of PPPoE "dumbbell MTU". This will reduce intra-LAN throughtput.<br /><br />o IPv6 firewall doesn't permit any other inbound protocols other than<br />TCP / UDP. At a minimum, other transport layer protocols such as SCTP or DCCP should be possible to allow through the firewall by specifying their transport layer protocol number, predefined names for them would be better. An option to permit all transport layer protocols should also be provided if the user wants to have all their IPv6 firewalling performed by their host's IPv6 firewall.<br /><br />o ULAs don't include a 41 bit random number, it is set to all zeros.<br />
getting ipv6 into avm's firmware is all very well. but their ipv4 implementation is badly, badly broken because it wrongly assumes everything gets NAT'ed behind one public IPv4 address and inbound port forwarding can only go to one box on the LAN: ie incoming SSH (say) can only get forwarded to one internal host. i have no idea why they can't/won't fix this.<br /><br />avm's customer support is remarkably clueless too. they think everyone runs windows, even when you tell them you don't.