2012/07/01

IPV6: Adoption driven by smartphones?

Horace Dediu of asymco charts commodity computing up-take (as yearly sales) since 1975 in  "The evolution of the computing value chain". It includes personal computers, smartphones and more.

Current PC (desktop+laptop) and Android sales are about equal at 350MM/year, with iPhones ~150MM/year and iPads closing in on 100MM/year: or near 1Bn new personal compute devices per year.

With current growth rates, how soon will we saturate the IPV4 address space?


There are two questions in there:

  • How many IPV4 addresses are left?
    • Where? Each region has its own challenges.
  • How fast is the demand for Internet addresses growing?
    • New Sales are not the whole story: replacements account for most sales in mature markets.
      • The active PC fleet is now somewhat static.
    • When the smartphone and tablet markets mature, Sales will still hold up with replacements catering for the 2-5 year product life.
The 32-bit IPV4 address space can assign no more than 4Bn unique identities.
My guesstimate of realised assigned unique addresses is 2-3Bn. 

Smartphone network providers can use private addresses and NAT's to hide 10,000 "access-only" devices behind a single IP number, somewhat reducing the problem.

Looking at Horace's charts:
  • smartphone sales are doubling each year
  • the high-growth phase may last 5-7 years, but that's not data in the chart.
  • we can't know if there is a new App coming that will require fully exposed IP addresses, requiring IPV6.
The decider will be the Network Operators: If they see an advantage, marketing or technical, in going to IPV6.

Or maybe when total smartphone and tablet sales are well over 1Bn/year.

No comments: