4 ms·

IPv6 was poorly designed, and will never be fully adopted because of this. Unfortunately, we seem to be stuck between IPv4 and IPv6 when really, what we need is

by imdsm 5y ago

IPv6 was poorly designed, and will never be fully adopted because of this. Unfortunately, we seem to be stuck between IPv4 and IPv6 when really, what we need is a new solution, something better designed and easier to adopt.

camkego 5y ago
That's a strong statement about IPv6, is there an article or paper you are referring to?
- imdsm 5y ago
  Not really. Just personal experience. And I think that's the key point here - coming up with a design that is fit for the world (wide web), and not just academically viable. I suppose the evidence I would give to back up my statement is that if it was well designed, it would have been adopted. As it stands, I only find it is supported among the more distinguished services that I use.
  - vore 5y ago
    There's a difference between "poor design" and "difficult to adopt", and I think confounding the two doesn't make sense. For another example, Python 3 has much better design than Python 2, and the difficult adoption a result of having people fix what they had written using the poor Python 2 design, and making it easier to adopt would have meant preserving the poor choices. I don't know enough about IPv4 vs IPv6 to make a judgment in this particular case, but I don't think your argument broadly in itself here follows.
Plasmoid 5y ago
I've heard people complaining how hard IPv6 but I've never see a plan that is actually easier than migrating to IPv6. It's always massive CGNAT investment or making non-backwards compatible changes to IPv4 that is somehow "easier" to work with.
- kmonsen 5y ago
  That is like all the how to terraform Mars plans. I know what is easier than living on any other planet, saving the one we are currently on right now. That being said I think it is good people are so good at being non-conformist, at least if there was any other solution out there we would more than likely have found it by now.
  - est31 5y ago
    The comparison is wrong. ipv6 is already fully developed, while ipv4 is getting crowded. There are already more devices connected to the internet than there are ipv4 addresses. Already Android devices alone are more than 3 billion, PCs 1.5 billion. The limit is at 2^32 = 4.3 billion. It would be a better comparison if Mars were already habitable and one only had to move humans over. Because that's all we have to do to make ipv6 suitable.
    - nlitened 5y ago
      If I understand correctly, due to NAT we’re not limited by the number of computers vs available IP addresses, but rather by number of simultaneous connections vs (number of IPs) x (number of IP ports)
      - est31 5y ago
        You are right, NAT is why the internet still works in the first place. However, NATs have many issues, like severely harming peer to peer capabilities. There is also additional complexity. My ISP frequently has downtimes of their NAT infrastructure. Then there is the more complex routing tables. ipv4 is highly fragmented so users with many users don't have large contiguous blocks, but multiple smaller ones. This causes routing tables to bloat up, which increases routing overhead of the packets. Lastly, there is the ridiculousness of paying for ipv4 addresses in the first place. Ideally, companies could just register them, having to pay a small service fee, that's it. In ipv6 this is the case, but in ipv4 you see Amazon paying huge sums for large ipv4 blocks. More importantly though, this cost is also handed down to customers, at least for hosters like Hetzner. For small servers, this can quickly be a significant fraction of the cost.
nativeit 5y ago
Wouldn't this be the kind of thing that would be ripe for a startup to make easier/more accessible through abstraction? Maybe the risk of fracturing an already dense protocol into multiple competing approaches isn't worth it, but my personal experience with IPv6 is also that it can be intimidating and confusing for the uninitiated (of which I am).
sterwill 5y ago
IPv6 works fine. It's all over my house, my mobile devices use it all the time (they're probably IPv6-only over LTE at this point), and I don't have any problems when I occasionally have to type in an IPv6 address. I don't see how you'll wrench IPv6 out of the billions (?) of devices using it and make them use anything other than IPv4 at this point.
brian_cunnie 5y ago
> IPv6 was poorly designed. I'd like to point out some design decisions I love about IPv6: - 128-bit addresses instead of 32-bit. I can't imagine having a shortage of routable IP addresses for a long, long time. - I don't have to use NAT if I don't want because there are enough IP addresses to go around. We've grown so used to NAT that we take it for granted, not recognizing it for the ugly hack it is. On my home network, just about every machine has a routable IPv6 address. - A consistent subnet: it's almost always gonna be a /64. When I lay out my networks, I don't have to regret specifying a /24 and then, as the office grows, wishing I could extend it to a /22. Every IPv6 subnet is big enough. - A consistent subnet means that you don't need to use a subnet calculator as much. - Not wasting .0 and .255 addresses (network & broadcast addresses). It's particularly bad with very small subnets (e.g. an IPv4 /30 wastes half its usable IP addresses). - On that note, I like having my machines being able to have a .0 (::) address. My favorite IPv6 address? 2600::. You can ping it to see if your connectivity is working. - ::1 is loopback. 127.0.0.1 is lookpback on IPv4, but 127 is a kinda weird number (okay, it's 2^7 - 1, but it's still weird). - Without NAT I don't have to worry about network collisions when I VPN into corporate ("What? They're using 10.0.0.0/24? That's my home network's subnet! Dang!").

by imdsm 5y ago

camkego 5y ago

That's a strong statement about IPv6, is there an article or paper you are referring to?

imdsm 5y ago
Not really. Just personal experience. And I think that's the key point here - coming up with a design that is fit for the world (wide web), and not just academically viable. I suppose the evidence I would give to back up my statement is that if it was well designed, it would have been adopted. As it stands, I only find it is supported among the more distinguished services that I use.
- vore 5y ago
  There's a difference between "poor design" and "difficult to adopt", and I think confounding the two doesn't make sense. For another example, Python 3 has much better design than Python 2, and the difficult adoption a result of having people fix what they had written using the poor Python 2 design, and making it easier to adopt would have meant preserving the poor choices. I don't know enough about IPv4 vs IPv6 to make a judgment in this particular case, but I don't think your argument broadly in itself here follows.

Plasmoid 5y ago

I've heard people complaining how hard IPv6 but I've never see a plan that is actually easier than migrating to IPv6. It's always massive CGNAT investment or making non-backwards compatible changes to IPv4 that is somehow "easier" to work with.

kmonsen 5y ago
That is like all the how to terraform Mars plans. I know what is easier than living on any other planet, saving the one we are currently on right now. That being said I think it is good people are so good at being non-conformist, at least if there was any other solution out there we would more than likely have found it by now.
- est31 5y ago
  The comparison is wrong. ipv6 is already fully developed, while ipv4 is getting crowded. There are already more devices connected to the internet than there are ipv4 addresses. Already Android devices alone are more than 3 billion, PCs 1.5 billion. The limit is at 2^32 = 4.3 billion. It would be a better comparison if Mars were already habitable and one only had to move humans over. Because that's all we have to do to make ipv6 suitable.
  - nlitened 5y ago
    If I understand correctly, due to NAT we’re not limited by the number of computers vs available IP addresses, but rather by number of simultaneous connections vs (number of IPs) x (number of IP ports)
    - est31 5y ago
      You are right, NAT is why the internet still works in the first place. However, NATs have many issues, like severely harming peer to peer capabilities. There is also additional complexity. My ISP frequently has downtimes of their NAT infrastructure. Then there is the more complex routing tables. ipv4 is highly fragmented so users with many users don't have large contiguous blocks, but multiple smaller ones. This causes routing tables to bloat up, which increases routing overhead of the packets. Lastly, there is the ridiculousness of paying for ipv4 addresses in the first place. Ideally, companies could just register them, having to pay a small service fee, that's it. In ipv6 this is the case, but in ipv4 you see Amazon paying huge sums for large ipv4 blocks. More importantly though, this cost is also handed down to customers, at least for hosters like Hetzner. For small servers, this can quickly be a significant fraction of the cost.

nativeit 5y ago

Wouldn't this be the kind of thing that would be ripe for a startup to make easier/more accessible through abstraction? Maybe the risk of fracturing an already dense protocol into multiple competing approaches isn't worth it, but my personal experience with IPv6 is also that it can be intimidating and confusing for the uninitiated (of which I am).

sterwill 5y ago

IPv6 works fine. It's all over my house, my mobile devices use it all the time (they're probably IPv6-only over LTE at this point), and I don't have any problems when I occasionally have to type in an IPv6 address. I don't see how you'll wrench IPv6 out of the billions (?) of devices using it and make them use anything other than IPv4 at this point.

brian_cunnie 5y ago

> IPv6 was poorly designed. I'd like to point out some design decisions I love about IPv6: - 128-bit addresses instead of 32-bit. I can't imagine having a shortage of routable IP addresses for a long, long time. - I don't have to use NAT if I don't want because there are enough IP addresses to go around. We've grown so used to NAT that we take it for granted, not recognizing it for the ugly hack it is. On my home network, just about every machine has a routable IPv6 address. - A consistent subnet: it's almost always gonna be a /64. When I lay out my networks, I don't have to regret specifying a /24 and then, as the office grows, wishing I could extend it to a /22. Every IPv6 subnet is big enough. - A consistent subnet means that you don't need to use a subnet calculator as much. - Not wasting .0 and .255 addresses (network & broadcast addresses). It's particularly bad with very small subnets (e.g. an IPv4 /30 wastes half its usable IP addresses). - On that note, I like having my machines being able to have a .0 (::) address. My favorite IPv6 address? 2600::. You can ping it to see if your connectivity is working. - ::1 is loopback. 127.0.0.1 is lookpback on IPv4, but 127 is a kinda weird number (okay, it's 2^7 - 1, but it's still weird). - Without NAT I don't have to worry about network collisions when I VPN into corporate ("What? They're using 10.0.0.0/24? That's my home network's subnet! Dang!").