(Almost) All You Need To Know About IPv6 359
Butterspoon tips us to an article in Ars Technica titled "Everything you need to know about IPv6." Perhaps not quite "everything"; the article doesn't try to explain the reasons behind IPv6's meager adoption since its introduction 12 years ago. But it should be regarded as essential reading for anyone overly comfortable with their IPv4 addresses. Quoting: "As of January 1, 2007, 2.4 billion of those [IPv4 addresses] were in (some kind of) use. 1.3 billion were still available and about 170 million new addresses are given out each year. So at this rate, 7.5 years from now, we'll be clean out of IP addresses; faster if the number of addresses used per year goes up. Are you ready for IPv6?"
Re:May i be the first person to say (Score:4, Informative)
Re:Who's afraid of IPv6? (Score:3, Informative)
A very small peice of the IPv6's space is simply there to allow IPv4 to still work, so those devices won't have issues.
Besides, if everything else moves to IPv6, wouldn't that allow for IPv4 addresses to be freed up for this old systems?
~Francisco
Re:Running out of IPv4 (Score:5, Informative)
NAT though is NOT a solution, it's a patch, a fix to a problem of running out of space. There should be enough IP's out there for everyone, but the '/8 should be enough for the average company' idea from the 80's-early 90's screwed us all up. Each Coca Cola or IBM-owned computer for example could have it's own public IP, the way it should be, but they own 16M+ addresses, way too much for their needs. But anyway, IPv6 is going to keep us out of trouble for now until we make the same mistake (history has a tendency to repeat itself) and we have to invent IPv8 or so.
Next to that IPv4 has been missing some major features and runs into problems with large networks and (very) fast links (talking 10Gigabit for example) IPv6 will solve for us, it routes faster, it has inheritely support for multicast and jumboframes, IPSec and mobile versions while IPv4 usually has that functionality bolted on (sometimes implemented slightly different with each manufacturer).
Re:Address scarcity will not drive adoption of IPv (Score:3, Informative)
Re:Jumping on the bandwagon... (Score:3, Informative)
http://sonic.net/features/ipv6/ [sonic.net]
Or at least it's an IPv6 tunnel (not sure how that might differ from 'native').
I haven't got around to setting it up, but if/when I get my WRT54GL setup with OpenWRT I'll probably have it run IPv6 as well...
Re:All you need to know... (Score:5, Informative)
How to install IPv6 (Score:2, Informative)
Re:Running out of IPv4 (Score:5, Informative)
I suppose at that point, history will repeat itself and we'll have to invent IPv8.
Re:All you need to know... (Score:3, Informative)
IPv6 is way too painful (Score:4, Informative)
1) Our local provide (XO) doesn't even offer public IPv6 address space.
2) ARIN wants thousands of dollars PER YEAR for portable address space.
3) Identifying what/how-to use a substitute for the deprecated "site-local" addressing. Tracking this down took days of searching and piecing things together. All the docs agreed that site-local was deprected but rarely mentioned what was going to take its place. Here is some links to what was found, MS has surprising helpful documentation:
http://www.microsoft.com/technet/network/evaluate
http://book.itzero.com/read/cisco/0602/Cisco.Pres
Generate a global ID with either of the tools below:
http://www.kame.net/~suz/gen-ula.html [kame.net]
http://www.hznet.de/tools/generate-rfc4193-addr [hznet.de]
Additionally it is nearly impossible to control the allocation of hosts to specific suffixes. We often organize customers address space so that global catalog for each site are at, say,
In a nutshell, IPv6 tools and implementation on hosts fall far short of the enterprise tools used define and organize a LAN for IPv4 and until ease of use is at least on par with MS IPv4 DHCP point/click environment it is going to continue to languish. It absolutely must have integrated DHCP server redundancy with automatic failover/failback/sync so sorely lacking, LO these many years in MS offerings.
Re:What they DID leave out (Score:1, Informative)
Thomas Edison was a control freak and, from what I've read, an all-around asshat. Didn't stop him from being revered by the public and making millions on his inventions, many of which are still in use today, either in nearly their original form (light bulbs), or in modernized versions (movies, movie cameras).
Amazing how this shit continues to propagate. Edison didn't invent the light bulb, he (arguably) invented a better filament, and one which wasn't used much in the real world. By the time electricity was widespread enough to make large sales of bulbs practical, his version had long been surpassed.
After years of trying, Edison also failed utterly to invent the movie camera. However, he was paid a handsome sum of money to put his name on a camera and process invented by two brothers nobody has ever heard of in order to increase their sales (a ploy which worked brilliantly, showing that people were mindless morons back then as well).
It's pathetic the degree to which the average person is ignorant of the history of the technologies they use every day; Christ, I'll bet you think Marconi actually invented the radio, don't you?
Re:I'd have built our whole network on IPv6, but.. (Score:1, Informative)
Anyway, they tell you to get your IPv6 from your upstream because it doesn't screw the routing tables this way.
And, 2^16 is only 65k customers. It is not REALLY that big. If your provider doesn't have 65k end-customers (customers of customers, if they only sell to resellers), then, well..
FOR THOSE WANTING TO TRY IPV6 (Score:1, Informative)
A guide to easy obtaining an IPv6 tunnel is available at sixxs.net [sixxs.net], pretty simple and straightforward.
Re:I'd have built our whole network on IPv6, but.. (Score:3, Informative)
Re:Why it's not catching on (Score:1, Informative)
The addresses are 4 times as long, but the IPv6 header is 40 bytes instead of 20 for IPv4 (if there is no IP options).
There is no IP header checksum that needs to be recalculated for IPv6.
Most of the code that is executed during packet forwarding will be network driver code and route lookup (fast or slow path) which (if done correctly) is identical for IPv4 and IPv6 (except for longer keys for route lookup).
Sure, IPv6 extension headers can be expensive to process, but so can IPv4 options. But most packets will not contain extension header nor IP options.
The performance measurements I've done on packet forwarding on different stacks show a performance decrease from 1% to 10% for IPv6 depending on which stack/OS you use.
No longer BBN's (Score:3, Informative)
BTW the company changed its name to "BBN Corp." around 1995, at which time its commercial ISP operation took the name BBN Planet. That used Net 4, as well as ASN (autonomous system number, used by BGP) 1. In 1997, GTE bought them. In 2000, Bell Atlantic (l/k/a Verizon) took over, but as terms of the deal, BBN Planet became a separate partially-owned subsidiary called Genuity. It did an IPO and burned through billions in a hurry before tanking. Net 4 went with Genuity, and was acquired as a bankruptcy asset, with the rest of Genuity, by Level 3. But only the lower 1/4 of Net 4 -- the rest was already returned.
The other two Class A's were not for BBN's own use. BBN had run the ARPAnet for the feds, having built the first routers ("IMPs") in 1969. They were BBN's for government contract use only, and were returned to the assignment pool in the late 1990s.
Verizon kept BBN Systems & Technology, the non-ISP side, for a couple of years, but it didn't really fit. Eventually it was spun off to investors and BBN Technologies is again a separate company, mostly doing government research, and not an ISP. BBN's internal network uses a Class B (128.33, IIRC).
Re:All you need to know... (Score:3, Informative)
Or, in other words, it could feasibly be possible to never upgrade to IPV6.