11 ms·

So the largest expansion of general availability compute (and associated addressing needs) occurs and it ignores ipv6. Re: 50% ... Your ipv6 numbers are very l

So the largest expansion of general availability compute (and associated addressing needs) occurs and it ignores ipv6. Re: 50% ... Your ipv6 numbers are very likely mobile, are they not? And those all CGNAT to ipv4 land? So the entire ipv6 mobile space (the biggest victory of the ipv6 war) is basically behind the #1 thing that ipv6 hate more than anything: a NAT. If you can't get ISPs to support it............ ............... ............... ISPs! The central access points to the internet for consumers! THE MAIN PEOPLE THAT YOU SHOULD HAVE CORRALLED. If you didn't get these people on IPv6 within a decade of the protocol, well I don't know what to say. I can't explain the massive degree of failure. Well, I don't have to, look at this thread. I just hear excuses. Blah blah ISPs Blah blah VMs blah blah kubernetes. If the ISPs don't support ipv6, then the software applications CANNOT be upgraded. If it is 2022 and there are major failures in ipv6 support (and there are) in places like comcast and the like, then whose fault is that? It is the fault of the ipv6 organizations, designers, ratifiers, etc. To underline, in America we have MASSIVE monopolies on consumer ISP access, and have for decades. That means in order to wrangle these organizations, you, what, get 10 people in the room and the FCC and get it hammered out? IPV6 isn't just a failure of protocol design. It is a massive failure of organization and outreach. The attitude, rife throughout all ipv6 conversations, is the protocol designers forced it on the internet and said "take it or leave it". The issue is ipv4 has ALWAYS been easier for a programmer. ipv6 always makes for more headaches. This is a massive massive failure. IPV4 networking sucks! NAT! bridges! Gateways! Prozies! port forwarding! CGNAT! It's still all easier than ipv6. Think about that. The issue here is that ipv6 has no hammer. It didn't have the FCC on board or some ultra-powerful group that could deliver the hammer. This is like IE6 (what is it about the number 6? don't answer that), that required a grassroots rebellion in silicon valley to dislodge. Well, and Google Chrome. GET A HAMMER. Idea: GOOGLE IS THE HAMMER.

unquietwiki 4y ago
Windows Vista came out in 2007, and it explicitly had a dual-stack IPv4+IPv6 implementation. Around that same time, the BSDs were finishing up KAME, and Linux support was decently mature. That was 15 years ago. Software and OS-wise; with the exception of some features of Mikrotik routers + managed Meraki & Ubiquiti stuff just now getting onboard; the support is there. Node's using it properly now. I was able to update Pidgin to use it 10+ years ago. So I think we've got the software-side long figured out; we just needed to get the damn vendors on-board. I'm glad we agree on NAT sucking. Though given a lot of ISPs are doing DHCPv6 without fixed net-ranges to clients; a work environment does benefit from having ULAs still; sometimes even with NAT66. Performance & security-wise, IPv6 still wins that battle: most routers disable automatic inbound traffic, and you can pinhole traffic directly between hosts.
- BizarroLand 4y ago
  I'm willing to bet that every ISP in America has IPV6 capabilities but the ones that have not activated it have done so out of corporate lethargy. The bosses aren't yelling, "TURN ON IPV6 AND MAKE IT WORK" so the peons have no reason to flip the switches. The bosses have no reason to push for IPV6 as it will increase support costs and add work hours but it will not increase income or improve profits. We need a mandate from the FCC that says that any ISP that has accepted federal funds to build or deploy internet access systems must turn on IPV6 to the consumer or face penalties. Short of that, it's not going to happen.

unquietwiki 4y ago

Windows Vista came out in 2007, and it explicitly had a dual-stack IPv4+IPv6 implementation. Around that same time, the BSDs were finishing up KAME, and Linux support was decently mature. That was 15 years ago. Software and OS-wise; with the exception of some features of Mikrotik routers + managed Meraki & Ubiquiti stuff just now getting onboard; the support is there. Node's using it properly now. I was able to update Pidgin to use it 10+ years ago. So I think we've got the software-side long figured out; we just needed to get the damn vendors on-board. I'm glad we agree on NAT sucking. Though given a lot of ISPs are doing DHCPv6 without fixed net-ranges to clients; a work environment does benefit from having ULAs still; sometimes even with NAT66. Performance & security-wise, IPv6 still wins that battle: most routers disable automatic inbound traffic, and you can pinhole traffic directly between hosts.

BizarroLand 4y ago
I'm willing to bet that every ISP in America has IPV6 capabilities but the ones that have not activated it have done so out of corporate lethargy. The bosses aren't yelling, "TURN ON IPV6 AND MAKE IT WORK" so the peons have no reason to flip the switches. The bosses have no reason to push for IPV6 as it will increase support costs and add work hours but it will not increase income or improve profits. We need a mandate from the FCC that says that any ISP that has accepted federal funds to build or deploy internet access systems must turn on IPV6 to the consumer or face penalties. Short of that, it's not going to happen.