4 ms·
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
by mindslight 12y ago
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.
- icedchai 12y agoI set up a HE.net v6 tunnel over 4 years ago now. Sometimes, I ask myself why I bothered.
- mindslight 12y agoExperience 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.