Slashdot Mirror
Internet Groups To Stream Live IPv4/6 Announcement
revealingheart writes "On Thursday, 3 February 2011, at 9:30 AM Eastern Standard Time (EST) [14:30 UTC/GMT], the Number Resource Organization (NRO), along with the Internet Corporation for Assigned Names and Numbers, the Internet Society (ISOC) and the Internet Architecture Board (IAB) will be holding a ceremony and press conference to make a significant announcement and to discuss the global transition to the next generation of Internet addresses. We invite all interested community members to view the webcast of this event."
"Yes, we have no bananas."
"National Security is the chief cause of national insecurity." - Celine's First Law
No awareness of timezones whatsoever.
It's going to be 6:30 in California, people!
Sheesh, if you want people to watch your announcement live, you need to schedule it when as many folks as possible are AWAKE.
Well, NAT64 is like purgatory. You left earthly IPv4 in God's grace but you receive punishment for continued involvement with 32-bit sinners. And, yeah, I suppose purgatory needs a sunset period.
The shit did not hit the fan yet. No one really cares about IANA's pool running out -- but not being able to obtain them from RIPE will be a serious problem.
The creatures outside looked from Alt-Right to Antifa; but already it was impossible to say which was which.
No one really cares about IANA's pool running out
RIPE probably cares a lot about IANA's pool running out and they will surely spread all their troubles downstream to all of the account holders.
You missed a token ceremony of them handing out the last IPv4 pools to the RIRs. That's it. It was like a cheesy award ceremony
What's the point in streaming an announcement if NOBODY HAS AN IP? :P
Ahem. Some of us got an IP while there was still time. Of course, seing you have a 7-digit Slashdot id, I can see where you're coming from.
Bring back NAT-PT! It was prematurely obsoleted due to scalability concerns. Those concerns are indeed valid, but only for large networks. On a home network with a couple of users it is a perfectly viable solution. Put NAT-PT on a router appliance, give it an IPv6 address, and it will let the home network transparently pretend that IPv6 does not exist. Yes, there are a few obvious problems with the few protocols that send IP addresses, like bittorrent, but a simple client fix can easily send hostnames instead. Otherwise, it will just work, and nobody will have to care about IPv6 except ISPs.
Many of the transition problems arise from the insistence that everybody want IPv6. Normal people don't care about IPv6, don't want IPv6, and couldn't care less what it is. Instead of starting to convert from the bottom up, with users going IPv6 first on their home networks, and then the ISPs and backbones switching when everybody has moved, do it the other way around. Convert the backbones to IPv6 down to ISP level. Then the consumers can use NAT-PT appliances to pretend that that did not happen and keep on going without any disruption.
Everybody talks about how big of a pain in the ass IPv6 is going to be to deploy
The big problems are
1: the chicken and egg problem. While everyone can access v4 servers there is little point in adding v6 to them and while a susbstantial portion of servers are v4 only ISPs won't want to make users v6 only.
2: infrastructure equipment, older home routers often don't do IPv6 at all and some older proffesional routers only do v6 in software which is much slower than doing v4 in hardware.
3: application software support. To support IPv6 apps must do certain things differently from with v4, it's not hugely difficult to port an app if you have the source but if you don't it's an extreme PITA to hack support in.
Be honest, is this all another Y2K where everything is just going to be smooth and a bunch of idiots that aren't talented enough to do the work but smart enough to know there is some kind of an issue are going to scream about the sky falling?
What is almost certain to happen is that home lusers will be pushed behind ISP level v4 nat to free up IPs for more important/lucrative uses. They may or may not also offer IPv6. If you use anything other than traditional client apps (NAT traversal techniques get flakier the higher the load on the NAT and is incompatible with certain types of NAT so don't rely on apps that use it continuing to work) on a home connection you should expect to have problems and should press your ISP for information on their plans and/or make contingency plans for a situation where the only apps you can run on a home connection are traditional client apps.
Further anyone trying to expand their public IP networks is going to be in for a lot of pain. While ISPs will be able to provide IPs by recovering them from home lusers that is a lot of effort for the ISP and is almost certain to come at a price.
note: i'm known as plugwash most places but i screwd up registering that here somehow in the past and now can't register
See http://www.iana.org/assignments/ipv4-address-space/ipv4-address-space.xml
224/8 to 239/8 are actually used for multicast. We use them in our company for example (makes for *great* bandwidth saver for media streamings and stuff)
240/8 to 255/8 (so called E-classes in old days) however were "reserved for future use".
The problem with reusing them is exactly what you state -- way too many firewalls and routers around the world drops those, and so they are effectively unusable for global (Internet) routing - you would need to fix the whole world before you could safely use them (who'd want IPv4 adresses that won't work on 99% of the Internet?), and apparently fixing the whole world is much much more complicated than just deploying IPv6.
What you didn't know is that the presentation is a performance art piece. It depicted just about what will happen in a few months if you ask for a block of v4 addresses.
We likely won't need an IPv6 replacement until we've colonized most of the Universe. The IPv6 address space is 128 bits long, although the lower 64 bits are reserved for individual host addresses, so we could view the address space as being 65 bits wide.
That gives us 2^65 addresses, more or less.
2^65 = 3.68934881 Ã-- 10^19
So, that's 36.89 QUINTILLION addresses. That's really, really, a lot of addresses. To put this in perspective, according to WikiPedia, there's 100-400 Million stars in the Milky Way. If 50% of those stars had 1 colonizable planet, and we wanted to create an Intergalactic Internet (which, of course, would require faster-than-light comms, but, hey, this is just an example), so that we colonized 200 Million planets, each planet could have:
3.689 * 10 ^19 / 2 * 10^8 = 184 * 10^9 addresses per planet.
So, you get somewhere around 184 Billion addresses per planet. That's not an exact number, because, remember, I treated the address space as 65 bits. That's because you get a 64 bit network prefix, and a 64 bit host address. A network, however, can have more than 2 addresses (which, the simplification of saying we have 65 bits of addresses basically assumes an average of 2 hosts per network), so this estimate is probably actually low. On the other hand, allocation guidelines also 'waste' a lot of the address space (like saying that every household and small business should get their own /56 and larger businesses and organizations should get a /48, so that they can run their own subnets if they want), but we actually have plenty of address space to waste (no, really, we do - worst case scenario is that the IPv6 address space is basically 48 bits wide - that's still a LOT of network prefixes).