I don't have a good feel for the numbers involved here. Things like this always report: "We are out!... only 4,000,000 or so addressed remaining!" Four million sounds like a lot to me. Some number of the IP addresses that are counted as gone are also still unused.
Can someone give me some idea of the situation with numbers that are easier to understand? Can we still limp through for a few years, or is this actually the crisis point?
It does seem a bit anti-climactic. "We're out! We only have a /10 remaining, and we'll only be handing out /24 - /22 allocations!". Which does seem an awful lot like "flowing freely" (I'd be happy with just a personal /27 myself) until you put it in context with those early non-aggregable /24 allocations in 192/8 being generally viewed as a mistake for router memory, and the subsequent move to only give out portable allocations to larger service providers.
IMHO all of the IPv4 exhaustion articles are basically scaremongering attempting drive adoption ahead of the real pressing need. Which isn't the worst goal, as it encourages people to become familiar (have YOU setup a tunnel from he.net, sixxs, or with 6to4/miredo and started playing around?). But I suspect we'll be hearing variations on the same "we're out! (but not really out)" story for the next 5 years at least.
But besides a larger address space, v6 doesn't really give new features (besides plausible deniability of what constitutes a host address) and just adds to packet overhead. So the only compelling reason will be the ever growing number of endpoints that are v6 only, and behind NAT (from businesses who decide that providing v4 addresses to new customers by default is too expensive). Other businesses wanting to better serve/track those customers (avoiding NAT64 etc) will then finally have a real reason to treat IPv6 as a first class concern.
Experience and familiarity, mainly. When Linode enabled native v6, setting it up was a no-brainer. I'd also be comfortable getting a v6-only server if it fit my need.
Also it's a nice backup link for NATted WAN machines in case tinc/autossh/etc accidentally goes down.
> Since it began operating in 2002, the organization has assigned more than 182 million IPv4 addresses throughout Latin America and the Caribbean.
182 million addresses in 12 years is roughly 15 million addresses per year, or 1.25 million addresses per month. At that rate 4 million addresses would last maybe 3 months.
Yes, but they could have issued 30m each to 5 different ISPs that are only half way through their allotment. If I want a fixed IP at my house or my school I don't get it from the internet address registry, I get it from Boston University or RCN.
Can someone give me some idea of the situation with numbers that are easier to understand? Can we still limp through for a few years, or is this actually the crisis point?