SoylentNews is people
SoylentNews
from the don't-give-huge-blocks-to-businesses dept.
Things are looking up for our next-generation internet.
[...] But the shortage of IPv4 elbow room became a steadily worsening issue -- have you noticed all those phones that can connect to the network now, for example? So tech companies banded together to try to advance IPv6. The result: World IPv6 Day on June 8, 2011, when tech giants like Google, Facebook and Yahoo tested IPv6 sites to find any problems. For a sequel, they restarted those IPv6 connections and left them on starting on World IPv6 Launch Day, June 6, 2012.
Back then, there was still a risk that IPv6 wouldn't attract a critical mass of usage even with the tech biggies on board. The result would've been an internet complicated by multilayer trickery called network address translation, or NAT, that let multiple devices share the same IP address. But statistics released Wednesday by one IPv6 organizer, the Internet Society, show that IPv6 is growing steadily in usage, with about a quarter of us now using it worldwide. It looks like we're finally moving into a future that's been within our grasp since the Clinton administration.
"While there is obviously more to be done -- like roll out IPv6 to the other 75 percent of the Internet -- it's becoming clear that IPv6 is here to stay and is well-positioned to support the Internet's growth for the next several decades," said Lorenzo Colitti, a Google software engineer who's worked on IPv6 for years.
[...] How much room does IPv6 have? Enough to give network addresses to 340 undecillion devices -- that's two to the 128th power, or 340,282,366,920,938,463,463,374,607,431,768,211,456 if you're keeping score.
(Score: 4, Informative) by VLM on Thursday June 07 2018, @05:01PM
Yeah it don't work like that. The RIRs get subnet chunks like /teens (like /16, or maybe /12, or maybe /20) and they dole out space to ISPs in big ass chunks like /32 for a really big ISP, and the ISP customers are supposed to get a /48 or /56 or for the really cheap bastards a single nasty /64 to use as they please. A ipv6 lan is /64 nobody subnets smaller than that, not much, so a /56 seems extravagant for home user, but its only 8 bits of LAN/VLAN. I myself use about 10 VLANs at home for somewhat obscure reasons and I think it would be hard to use more.
Or rephrased, depending on how cheap / shitty your ISP is, you'll get somewhere from 1 to 16 bits of /64 sized LANs to use, which can go pretty fast if you're a big enterprise, and your ISP probably got 16 to 32 bits worth of those LANs to dole out.
You wouldn't believe how butthurt some ipv4 admins get over the idea of assigning a whole /64 to a pt-pt WAN interface. In ipv4 weird ass subnet masks were the rule but the first rule of ipv6 is the subnet mask of a network is ALWAYS /64 unless you're aggregating routes (aka your device is itself a router so its routing a /48 to multiple real world /64 physical interfaces, or at least VLANs). Everything in ipv4 is figuring out what 99.88.77.66 is WRT a /27 vs a /28 or slash WTF as a subnet mask, but every network in ipv6 is a /64 all the time...
Sorta like someone claiming in theory that given the (xxx) yyy-zzzz phone number format, we can have 1e10 addressible phones but it don't work that way in practice. Recently with LNP its not as ridiculous as it would have been in 1990 but whatevs, and nobody will ever have the 911 NPA or the 800 NPA etc...
Also note just like the old days of classful ipv4 (listen to the real dinosaurs roar an elderly groan) there sure as hell are not 128 bits worth of assignable ipv6 addrs. I wonder how many morons have copied 2001:db8:: out of an instruction manual and still don't know why it doesn't work.