One week Google stats floor above 40%

[Mark12547]

Looking at Google Stats for the period of June 10-17, 2003, we went a whole week above 40%, which is a good sign!


We had a previous week period above 40%, December 24-31, 2022, but around Christmas we know that, at least in the United States, many people are off at least several of those days and are more likely accessing Google from their residential ISPs than from work or school and, at least in the United States, residential ISPs are more likely to have dual stack than corporate networks.

[Mark12547]

In other news, it looks like the auction prices of IPv4 addresses have been trending down, about 25% drop in the past six months.

[K6USY]

Wonder how long it will take us to get to 50%, and will that be a tipping point.

[Mark12547]

Wonder how long it will take us to get to 50%

A linear projection from the troughs (weekly lows) over 4 years (24.53% on June 17, 2019 through 40.47% on June 20, 2023) gives me November 11, 2025. Since adoption curves tend to be an S-curve overall but could have lots of noise, I would expect the troughs in Google's IPv6 adoption chart to reach 50% in 4Q2025, but not necessarily in November 2025. But predictions are difficult, especially about the future.

and will that be a tipping point.

I wouldn't expect a "tipping point" as such, but more likely a continuation of the S-curve, such as explained at https://www.stratechi.com/adoption-curves/

Outside factors could disrupt the curve, such as some governments (including the United States federal government) placing IPv6 as a requirement in their RFPs and some countries requiring ISPs in general to support IPv6.

If the federal government started having their public-facing websites be IPv6-only, that would apply pressure on ISPs to have IPv6 (as well as IPv4) so the public could interact with those websites, and that would accelerate the adoption process faster among ISPs than the S-curve would suggest, but may end up having little effect on corporate networks that often have legacy components or the decision makers may not see a need to upgrade to dual-stack.