Hacker News Reader

Comment by ghoul2 | Hacker News Reader

This is required by law. [0][1] Every ISP is required to be fully ipv6-enabled by the end of this year. The deadline has been pushed back multiple times , most recently due to covid. But this time it seems its going to stick. All lagging ISPs seem to have deployed v6 at this time.

There has been a "National IPV6 Plan" in place since 2010, and full coverage has been pushed under that plan in stages. Anyone interested can read all about it in (extensive) detail[2].

0. https://dot.gov.in/ipv6-transition

1. https://dot.gov.in/sites/default/files/Revision%20in%20IPV6%...

2. https://dot.gov.in/sites/default/files/2016_11_18%20IPv6%20N...

View on HN

Why is it so hard to pull off the same regulation in other parts of the world? It isn’t like supporting IPv6 is black magic…

For the US, if I had to guess, ISPs have huge lobbying presence in Washington. Of course they won't want to make more work for themselves.

CGNAT is pretty expensive compared to rolling out IPv6 and tunneling IPv4 on it using stuff like MAP-E. I guess they still have no need to NAT everyone ATM because they have plenty of IPv4s to distribute to their subscribers.

That is correct. My ISP in Ohio gives me both IPv6 and global V4.

CGNAT is pretty much the only setup on mobile though. Dunno if any equipment vendor doesn’t do that.

Would worry about most average cell phones sitting directly on the public internet.

I think there’s a reason we haven’t heard about mass-owns of mobile phones and it’s because they’re always effectively behind firewalls, not because of their superior security.

This is not correct. Especially in the USA lots of mobile phones now have IPv6 connectivity.

From https://www.arin.net/blog/2020/01/16/mobile-edge-of-the-inte...

"80% of smartphones in the US on the major cellular network operators use IPv6 and major mobile networks are driving IPv6 adoption with Verizon Wireless at 84%, Sprint at 70%, T-Mobile USA at 93%, and AT&T Wireless at 57%"

And this were numbers from four years ago.

NAT is not a firewall and it's dangerous to think of it as such, because it causes people to follow a lax approach to security and think they are protected, when they actually aren't.

That’s fine and all: but are they issuing mobile phones a publicly addressable IP or not? Can a Verizon subscriber run a web server on their mobile phone and share their IPv6 or not?

NAT66 is, in general, utterly broken and most OSes prefer IPv4 to ULAs - thus, if a mobile subscriber can connect to an external IPv6 service, it's very likely the address they were given by their carrier is publicly routable.

From practical standpoint, most common NAT setups come with "default deny incoming" firewall, and systems have relied on this fot a very long time.

I hope that IPv6 is rolled out to homes, most ISPs would do the right thing and set up customer routers firewall in "default deny incoming " mode. Otherwise, the number of worms would increase a lot.

We may argue that Windows Firewall (and macOS) by default already do that. On Linux, I usually set nftables like such on all my machines, but it's trivial for "user friendly" distros to ship sane defaults for nftables.

What scares me a lot is that lots of folks out there still rely on iptables, and do not realize that they need ip6tables to set IPv6 tables - I dread of the moment some random ISP will push IPv6 routes, SLAAC picks up an IPv6 on a random Linux box and that box gets instantly pwned because it has no ip6tables rule whatsoever blocking inbound traffic.

The many embedded devices in homes are much scarier than home sysadmins not configuring their boxes correctly.

For some reason CGNAT seems to be the norm here in Sweden, IPv6 is still almost nowhere to be seen, only existing in a select few cities.

I've been behind CGNAT since I think 2010, at first with DSL with a small ISP and then with one of the big ones with fiber (bredband2), who introduced CGNAT many years ago to prepare for moving to IPv6, which they have still yet to do.

They even stopped supporting their 6rd tunnel apparently.

India has lots of people and not much IPv4. The US has lots of IPv4 and not nearly as many people.

Is it the ISP's or holders of address blocks?

I wonder how much all the IPv4's on Earth are worth.

They apparently go for 20-60$/IP. That can tell us something about why large IP holders don't want IPv6, it would reduce their 'investment' to near 0.

Is the reason this regulation doesn’t exist elsewhere because it is difficult?

Maybe other places just don’t have a desire to regulate this.

Obviously they will need to sooner or later. IPv4 address space is mostly depleted; at one point or another, regulation is the only thing that will help.

It is possible to use IPv6 without being compelled by law.

Because generally it would be a bad idea for the government to force certain technologies upon the private sector, even if in this case it might be fine.

Forcing micro USB and now USB-C on electronic devices has been one of the best ideas by the EU overall. Remember using ten different charging cables? Gone, due to a helpful piece of regulation by the government.

I still have the same issue as before except I’ve added a suite of USB C connectors. I still have micro/mini and regular USB devices on top of all the USB C stuff. Not to mention the premium for a USB C charger.

You have to admit it's a better connector mechanically, though. Also, the vast majority of USB-C devices charge fine from USB-A. The devices that require an expensive charger are things that have previously required a proprietary charger, like laptops.

> the vast majority of USB-C devices charge fine from USB-A

I’ve gotten myself stuck a few times with this assumption. My MacBook Air (and probably any MacBook) will accept a trickle at 5V, but my Dell laptop will not.

> Not to mention the premium for a USB C charger.

Only if you want fast charging. I run my MacBook Air over a USB-A to USB-C cable and it slowly charges under light use at around 2.5A and 5V. And tops off to 100% while not in operation. If you want more voltage, then yeah, you’ll need a more specialized charger, whether USB-C or proprietary.

Until someone updates the USB-C or PD standard.

Have you tried Nintendo Switch's USB-C charging? It is non-standard.

My 45W lenovo usb c charger works very well to charge switch, but the switch included charger wont charge anything but the switch.

It’s not, that was debunked.

It isn't a good policy and if that is one of the best ideas of the EU then they have failed as a political project.

1. Corporations are exactly the tool we use to work out what trade offs of cost v. standardisation are best. Everyone is going to claim victory for a bureaucratic committee then a few years down the track we'll discover technology moves on and the regulation is force people to make stupid choices. Removing the flexibility to adapt at the forefront of technology is going to turn out to be foolish in hindsight. These devices aren't even 20 years old and the EU has already decided to start fossilising the tech.

2. There is a decent chance that this will corrupt the USB standard itself. We're already seeing the bizarre profusion of naming [0]. If there is an incentive for corporations to create incompatible things that are all named "USB-C" they may well manage it. Calls to mind the OOXML vs ODF format wars.

3. This is a total waste of political attention. I suppose now we know how the EU ended up in such a catastrophic posture with their energy policy. Their best and brightest were back-seat-driving iPhone design. Thank goodness for the Europeans, they may not have reliable power but if they did they won't have to search around for the right cable. And if we need to send phones into the land war, we know what adaptors to send with them!

[0] https://en.wikipedia.org/wiki/USB#Naming_scheme

There are plenty of other good ideas coming from European unification. But standardizing the connector has certainly been a success. We already have 11 years of hindsight. And despite agreeing on micro USB initially, this hasn't stopped a smooth upgrade to the logical successor, USB-C.

The USB standard is messy, but I think there's been ample evidence that this is not because of the EU. Bluetooth is also very messy.

Point three is just nonsense, I got nothing to add to that.

> Thank goodness for the Europeans, they may not have reliable power but if they did they won't have to search around for the right cable

yeah... about that.... we in fact have electricity in the outlets, we have countries where the grid has lower chance of a power outage than the chances of a UPS failing. We dont have people rushing to buy gasoline for their generators as soon as a wee bit of bad weather hits..

hows it going in california right now? Babylon bee had a pretty funny, atleast from an outsiders perspective, "news" article:

https://babylonbee.com/news/california-man-takes-his-fiance-...

Except for Apple, all phone makers managed to agree to the USB-C standard without any government regulation.

What you're referring to is a very recent law (months old) that will take effect in 2024.

And it could limit future newer ports that are technologically superior (like lightning was when the standard was micro USB)

This is one of the situations where the proliferation of standards is preferable to technological superiority. I far prefer USB-C to be ubiquitously available over having marginally higher transfer or charging speed.

In this case, it seems different because IPv6 is the kind of thing that I doubt any free market competitor would ever be able to supplant. Part of that is the nature of IPv6 itself, as it was designed to be as future proof as it could be. I just don’t see a competing protocol as offering any kind of substantial benefit.

Indian isp are not using IPv6 though they seem to assign most users local Nat addresses my guess is looking how cheap internet is in India using real IPv6 addresses would increase internet prices

Can I ask you why? What makes using ipv6 more expensive?

I have no basis for this, but I would have to imagine router tables are significantly larger with IPv6, hence higher costs due increased complexity.

It's intended to be the opposite with fewer IPv6 prefixes needed, compared to the level of fragmentation for global IPv4 deployments combining many smaller prefixes to scavenge enough address space.

IPv4 930k -> 515k aggregated https://www.cidr-report.org/as2.0/

IPv6 161k -> 86k aggregated https://www.cidr-report.org/v6/as2.0/

Both of those are nothing compared to the requirements for CGNAT, tracking each individual TCP session.

Ah, but tracking each session is needed for censorship purposes no?

Because there is no significant benefit to the consumer

It has: significantly cheaper internet access once providers have to charge extra for expensive v4 addresses, complex NAT solutions and gateways.

Significantly cheaper is a bit of an overstatement.

Savings are pretty much capped at one month of internet service, as that’s roughly what an IPv4 address costs.

8.33% is not a significant gain to you?

Not really, especially as it’s not 8.33%, since it’s not a yearly cost, but a lifetime cost.

We had the same in Brazil, scheduled per connection type (DOCSIS was first, etc). It's the only way IPv6 will happen, and more people should push for this.

Yeah, looking at you Verizon FIOS.

It seems it was in plans for a long time undetected, but then they committed to enforcing it when Jio took off.

Reply on news.ycombinator.com