←back to thread

The IPv6 Transition

(www.potaroo.net)
215 points todsacerdoti | 2 comments | 20 Oct 24 05:54 UTC | HN request time: 0.406s | source
Show context
commandersaki ◴[20 Oct 24 06:57 UTC] No.41893472[source]▶
> This is the same as looking at a linear trend line placed over the data series used in Figure 1, looking for the date when this trend line reaches 100%. Using a least-squares best fit for this data set from January 2020 to the present day, and using a linear trend line, we can come up with Figure 2.

> This exercise predicts that we’ll see completion of this transition in late 2045, or some 20 years into the future.

Anyone willing to place a bet on this?

> While the design of IPv6 consumed a lot of attention at the time, the concept of transition of the network from IPv4 to IPv6 did not.

> Given the runaway adoption of IPv4, there was a naive expectation that IPv6 would similarly just take off, and there was no need to give the transition much thought. In the first phase, we would expect to see applications, hosts and networks adding support for IPv6 in addition to IPv4, transforming the internet into a dual stack environment. In the second phase we could then phase out support for IPv4.

I really don't understand this, how do you not make a transition plan the #1 requirement for selecting the next IP. (But the article goes on to say...)

replies(2): >>41898002 #>>41907333 #
kortilla ◴[20 Oct 24 19:49 UTC] No.41898002[source]▶
> Anyone willing to place a bet on this?

Ill bet against it. The tail on this one is going to be super long.

There are embedded systems today that are shipping in things expected to last 30 years with IPv4 only.

The logistics of the bet are going to be hard. I do see a world where IPv6-only becomes the default for ISPs and IPv4 becomes an add-on you pay for either from your ISP or from another via a tunnel. Does that world mean v4 is dead yet?

replies(1): >>41899840 #
1. tialaramex ◴[21 Oct 24 01:08 UTC] No.41899840[source]▶
The long tail doesn't matter. Once IPv4 traffic is a small fraction, the big transit providers will make it cost too much to bother with, and their customers (retail ISPs) will just cut it.

Only global IPv4 matters. If in fifty years there's still a device that insists on speaking IPv4 with the address 10.20.30.40 that will still work and it still won't matter to the Internet any more than it does now.

The appropriate comparison is leaded gasoline.

In my country this was never formally banned. You can't buy a new car which consumes it of course, they banned that, but the fuel itself is legal and for a while enthusiasts would travel to a retailer which still sold it, there might be one in the next town, or the next. Of course with fewer customers the price went up, further reducing customers and squeezing more retailers out, soon enough you might have an hour's drive to buy fuel. The wholesalers were next, if you sell a tanker of ordinary unleaded every five minutes, and a tanker of "high performance" unleaded every hour, why bother making the leaded fuel that shifts only one tanker per week across the whole market? It's not even worth reconfiguring your mixers to make it. So you mark it "No longer available" and gradually across the market the retailers can't buy more and there is no more leaded gasoline.

You can make your own leaded gasoline, but the volumes involved mean it no longer makes any meaningful difference, you could make your own lead paint too, if you're crazy, it doesn't make a noticeable difference to the world.

replies(1): >>41912231 #
2. kortilla ◴[22 Oct 24 08:08 UTC] No.41912231[source]▶
> The long tail doesn't matter.

We’re talking about the logistics of a bet, it’s the only thing that matters in this context.

The big transit providers pay very little to carry ipv4 prefixes. They will never even consider cutting it as long as there are any semi large content providers offering v4. Transit cores are prefix-free with segment routing so the cost is basically just the device that peers at the exchanges with other peers.

> The appropriate comparison is leaded gasoline.

It is not. The price dynamics make it so cheap to keep supporting ipv4 that it’s nothing like the unleaded switch. That’s before you even consider how dumb “banning the sale of new IPv4 devices” is.