Hacker News Reader

Comment by b112 | Hacker News Reader

That would be an incredible overreach for the EU, beyond anything it's ever attempted before. It would also be directly interceding in engineering standards, without there even being an overall, enforceable standards body (RFCs aren't mandated standards).

Really the dollar speaks here. If ipv6 is a real requirement, than companies not supporting it will be hurt fiscally.

Taking a step back, one might consider how incredibly... unwanted and unwieldy ipv6 is. It's just a pain to deal with. Pundits will say "Oh, it's easy!". Great. Now explain why something touted as utterly simplistic and with supposedly zero issues rolling out, and is cited as having immense upside, has its adoption fought by companies which have top tier talent working for them.

Sure. It's all that top tier talent that's wrong. They're confused. It's not ipv6.

View on HN

The Czech government mandates IPv4 to be turned off by 6/6/2032 for the official institutions and it will probably involve state companies too. https://konecipv4.cz/en/

IPv4 hurts innovation by basically charging for numbers more than an administrative fee hindering newcomers to enter the market. Everybody suffers a little because of the networking effect. However, IPv4 seems to be a pain for the big companies too. Some, like e.g. Facebook are really pushing it.

With the supposedly "top tier talent" why does e.g. Azure have such a bad IPv6 implementation? Why is Cloudflare or big parts of GitHub IPv4-only? Isn't that meant for developers? What about e.g. Twitter?

The biggest problem is not the infrastructure. All the important programs like DNS servers, operating systems and hardware do support IPv6 more or less for the last 10-15 years well enough. The problem is developers that are stuck in IPv4 mentally, doing GUI windows, that only accept IPv4, only testing on IPv4, specifying database columns to hold only IPv4 etc.

It's because they do not reap the benefits of ipv6 nor do they bear the burden of continuing to use ipv4. Those sites are big enough that they can simply rely on the fact providers ensure their customers can reach those sites. They could be using the OSI protocol and providers would still find a way to tunnel their customers to those servers.

The cost of ipv4 is ultimately being paid for by customers in the form of higher fees for their internet subscriptions. And this is exactly a case where the government can help the market operate more efficiently by demanding standardization.

I don't believe it's a talent/skill issue. It's a business issue. IPv6 is easy to deprioritize in the daily course of business because there remains a working alternative.

Is there really something wrong with IPv6? Isn't this just another case of resistance to change and the envisioned cost of it vs the benefits? I think it's just a matter of (even more) time.

Only regulated markets can be free markets. I don't really want to see more pathological cartels forming around internet access.

Amen, brother. Fact is, all critical limitations with IPv4 were hacked around by necessity, so the average user doesn’t care.

The main failure case is that it’s a painful or possibly impossible to setup your own public internet server from home, but ISPs are probably happy about that and games have figured out a work-around.

If IPv6 was a project at work, it would probably get killed as not a priority.

Reply on news.ycombinator.com