Champagne anybody?

[tiermutter]

Why have a VPN on an inherently unsafe platform?

I need VPN to talk to my local network at home and also my workplace.
For both me use v6 in first instance and v4 as failover (4G / DSL).
Where is there "an inherently unsafe platform"?
Where are the "safe" differences to v4?
 

[Greg_E]

All I read here can be summarized: ipv6 helps nobody to solve a problem, but results in signifcant problems. NAT is for me a feature, not a bug.

So: ipv6 deserves to die quickly. and for ever...

That's a -1

Me really happy to have v6 for VPN purposes since I am behind CGNAT.
For sure me can pay high amounts of money to get a public IP... but why, thanks IPv6?

My IPv6 at home is still not a public IP, I'm also behind CGNAT. I guess this would be an example of an ISP messing things up. I wish I had any form of public and static IP at home, would make several things much easier.

[verfluchten]

ipv6 deserves to die. no solutions, only problems.

IPv6 should have never been born, to begin with.
Instead of adding 1, okay, maybe 2 extra bytes to IPv4 and still keeping the IP address readable and memorizable, they bumped bytes up to words and added 2. This gives us more possible IP addresses than protons in the universe. Audacity? Nothing but.
A 5-byte IPv5 for lack of better idiom would have afforded us with more than 1 trillon addresses which is enough for all consumer goods on our planet, assuming that each has to be tracked. A 6-byte address would have been enough to track every potato. But that was not good enough for cowboys, so they doubled the width and made addresses impossible to remember. Isn't that genius?

[Patrick M. Hausen]

The width was quadrupled. And for sound technical reasons. Tell me you don't know networking without telling me you don't know networking, much?

I'm getting snarky so I better pull out of this discussion. About everything about the design of IPv6 is superior to the decades of accumulated band aid over band aid that is the present state of IPv4.

NAT is a terrible idea and violates the fundamental design principles upon which the Internet was built.

[verfluchten]

When bytes become words, it's doubling. That alone would have been enough, but no, they added more words. Now, try to remember a handful of IPv6 addresses for several years.

[bimbar]

I'm ambivalent. Maybe I know too much.

I think if everybody adheres to the standard, and a few newer RFCs are universally implemented, it's fine. There are still a few scenarios left where it doesn't work.

[verfluchten]

I eagerly admit that I don't know Jack S!
All I know is that an instance of IPv4 is a 32-bit integer that obeys the respective arithmetic and needs simple GUI controls.
An instance of IPv6 is a 128-bit integer that is not immediately supported on many platforms and requires way more complex GUI controls. It also takes 4x the space in memory and logs.
As an admin, I hate IPv6. As a dev, I hate IPv6. Had they simply added 1, 2, 3, or worst case scenario 4 bytes to IPv4, it still would have been a 64-bit integer that most platforms support today, and it would have solved the problem of too few IP addresses, without opening another whole can of worms.

[Monviech (Cedrik)]

They wanted to solve it once and for all with IPv6 for when we become a Kardashev Type II civilization.

Imagine how hard it would be a few thousand years from now on when we have passed the great filter and do interplanetary travel between different galaxies.

Try to hammer IPv12 through the space senate. 

[bimbar]

I eagerly admit that I don't know Jack S!
All I know is that an instance of IPv4 is a 32-bit integer that obeys the respective arithmetic and needs simple GUI controls.
An instance of IPv6 is a 128-bit integer that is not immediately supported on many platforms and requires way more complex GUI controls. It also takes 4x the space in memory and logs.
As an admin, I hate IPv6. As a dev, I hate IPv6. Had they simply added 1, 2, 3, or worst case scenario 4 bytes to IPv4, it still would have been a 64-bit integer that most platforms support today, and it would have solved the problem of too few IP addresses, without opening another whole can of worms.

As a dev, what are you using IP addresses for. Don't you have gethostbyname()?
As an admin, you should use your DNS server.
IPv6 has been developed to address a lot of shortcomings of IPv4, not only, maybe not even primarily the lack of IP address space.
Some of those are of course now obsolete, but there are useful attempts to solve real problems in there.

As for the OP, the direct quote seems to be that he does not see IPv6 replace IPv4 COMPLETELY.
I don't either, but that's not really a strong statement.
I still see room for IPv4 in LAN network addressing, though ULAs would be better. There will also always be public IPv4 connectivity for the big internet services, and IPv4 CGNAT will be with us for quite a while.

[Patrick M. Hausen]

Try to hammer IPv12 through the space senate. 

The senate is dysfunctional. Let's hand emergency powers to the supreme chancellor! 

[chemlud]

...too much champagne now? :-O

[Greg_E]

CGNAT needs to die!  (because it bugs me right now)

[verfluchten]

As an admin, you should use your DNS server.
IPv6 has been developed to address a lot of shortcomings of IPv4, not only, maybe not even primarily the lack of IP address space.
Some of those are of course now obsolete, but there are useful attempts to solve real problems in there.

As for the OP, the direct quote seems to be that he does not see IPv6 replace IPv4 COMPLETELY.
I don't either, but that's not really a strong statement.
I still see room for IPv4 in LAN network addressing, though ULAs would be better. There will also always be public IPv4 connectivity for the big internet services, and IPv4 CGNAT will be with us for quite a while.

None of this argues in favor of the current design of IPv6. You have not articulated any benefits of 8x 16-bit integers over 5-8x 8-bit integers that would have solved the problem of IPv4 being too few for Earth's population.

[Patrick M. Hausen]

64 bits is not nearly enough for route aggregation and hierarchies. While it is next to impossible to exhaust 2^64 numbers, it is entirely possible to exhaust 64 bits.

That's the reasoning behind the prevalent 64+64 (or 8+8) design.

It's all explained very well in Benedikt's presentation from 2018:
https://ripe77.ripe.net/archives/video/2287/

Also "running out of addresses" is by far not the only motivation for IPv6. SLAAC is pure genius. As Clemens Schrimpe uses to put it: "Wie macht man IPv6? An!" (How does one do IPv6? Switch on!)

Unless your provider is run by morons. About everything about the design is simple, auto-configuring etc.

I am running a 100% dual stack ISP and hosting service for over a decade. Just follow the RIR recommendations and it is dead easy.

[verfluchten]

64 bits is not nearly enough for route aggregation and hierarchies.

I dare you to demonstrate that.
If you are not a software developer, it is hard to appreciate the true magnitude of a 64-bit number.
It gives you another whole range of IPv4 on top of itself. There is no way in hell this is not enough.