Slashdot Mirror
China Lights Pure IPv6 Network
plui writes "An all IPv6 backbone was launched this weekend in China. 'CERNET2 is the biggest next-generation Internet network in operation in the world and connects 25 universities in 20 cities. The speed in the backbone network reaches 2.5 to 10 gigabits per second and connects the universities at a speed of 1 to 10 gigabits per second.' Here is a link to the story in the English version of China Daily, the online news site in People's Republic of China."
Heresy! I shall report you to the Department of Moral Standardization, the Bureau of Political Homogenization, and the Office of Internet Regulation!
But seriously, do you mean through proxies? Don't you worry that you may be caught / ratted upon?
A blog like any other.
Can't wait when Google will be available from my IPv6-only network.. ;-(
A few months ago, I asked a representative from Southwestern Bell when they would be ready to switch to IPv6. They indicated that they weren't even pursuing the issue.
I look forward to IPv6 just because it will kill the random port scanning by all the Windoze worms.
If we had already moved to IPv6, Code Red might still be looking for the second computer to infect.
Nice, yet another backbone that supports IPv6.
The only question is how to connect to an IPv6 network from behind a NAT based router. Connecting straight to the internet without the router I am able to establish a 6to4 tunnel no problem, but with the router in the way it is not possible. I am using a Linksys Gateway/Router.
As far as I know, porn is delivered also outside the internet, for instance in form of magazines. And no, you don't have to be literate to read those magazines... in fact, pregnant women are standing all over the country side selling those lewd magazines and video tapes. Pregnant women, because they can by law not be detained for this minor crime.
e nt_1461373.htm
Regarding netizens and their porn usage:
http://news.xinhuanet.com/english/2004-05/10/cont
From the article: One big benefit of the IPv6 is to solve the problem of shortage of IP addresses. In the current Internet based on IPv4 technology, the United States controls 74 per cent of 4 billion IP addresses, while the amount that China has is only equal to a campus of the University of California, despite its 80 million Internet users.
Although people think that with NAT all IPv4 related problems could be solved here we see a very good reason why the rest of the world could use IPv6. Most of the IPv4 ranges are in the US. The rest of the world just has to get by with whatever is left (Big companies gobbling up entire classes of IP ranges which they never really use should be obliged to gives those ranges back so others can use them).
Would this step be beneficial to the transition to IPv6? With the advent of the internet in other countries then the western world it could well be that things need to be sped up so that we will not see different internets. Has anyone done some real research on this subject? I know i keep hearing that with NAT and similar technologies IPv6 might not be necessary but is that really so given the rise of internet usage in Asia and other countries?
I moved to China less than 6 months ago - like any move, it is a combination of good news and bad, but the good news, like this, makes me happy I'm here :)
Add in the free HDTV set top boxes, all new subways in this province and living like a king for pennies a day, and the US won't see me back any time soon.
Five years ago when I worked for a major backbone provider (Qwest Comm.) and as I was talking to one of the global network engineers, he was wearing a shirt that had a Qwest logo and an IPv6 network under it. I asked him if that was for Qwest's IPv6 netblock and he indicated yes. We talked for about 15 minutes during which time he explained that Qwest, along with others like MCI/Worldcomm and AT&T (and several other big names you'd recognize if I cared enough to list them all), had a "parallel" IPv6 network running across the U.S. and some of them were even passing live internet traffic over these routers using encapsulation. However, the routers knew each other by their IPv6 address space.
So, five years later, how can the poster say CERNET2 is the biggest IPv6 network in the world? I would say "prove it". I would think several large international backbone providers who had links with each other passing live traffic would be considered pretty damned big!
- The shortage of GUAs (Global Unicast Addresses) is only half of the problem. The other part is route fragmentation. The defaultless routers (Those routers that route your packets safely through the backbone mesh) have a routing table of about 60.000 routing entries. It is a bit like fragmentation of a 96% full harddisk. The shortage is not only limiting you on placing new files, but it is also causing the file system to be inefficient.
- Let us first make clear, that there today is only a plan for about 15% of the address space, so before anybody starts rabbling, 85% is not only not in use, by is not even assigned a purpose.
- The IPv6 addresses has, like IPv4, a net part and a host part. Where this is variable in CIDR IPv4, it fixed in IPv6 to 64 bits. Every produced and future ethernet card will fit into one big flat switched net. Besides the full 48-bit MAC address, there is 16 bits for the EUI-64 extention which will be applied when new generations of network links are put in production.
- The largest part of the IPv6 address space is used for AGUA (Aggregatable Global Unicast Addresses); it is the 12.5% of the planned address usage. It is called so, because the allocation is done like CIDR addresses, based on where the main connection is routed to. There is reserved 13 bits (8192 combinations) for TLAs (Top Level Aggregators) which is large backbone providers like UUnet, PSInet and others of this size and extension. Each TLA slot will provide 24 bits (16 mill. combinations) for handing out to NLAs (Next Level Aggregators) in how many levels the business is shared (large ISP selling to small ISP). The customer will get 16 bits (65536 combinations) as a SLA (Site Level Aggregator) to make subnets with (You will get more subnets than you ever could dream of getting IPv4 individual address today). On top of this there is 8 bits reserved for future needs between the TLA bits and the NLA bits, but since we don't know where the shortage may be, decissions can be applied later.
- There is also reserved an address range for private address space. LLAs (Link Local Addresses) are auto configured addresses given to each interface as it is activated. There is also SLA (Site Local Addresses) which more ressembles the private address ranges (10.0.0.0, 172.16.0.0, 192.168.0.0) and is assigned by a local adminitrator's discretion. This about a half percent of the address range, but in practice is not possible use more than a fraction of it, for other reasons.
- Multicast addresses is also assigned a range of about a half percent, and also is for practical reasons only useable in a fraction of the space.
- Adding compatibility addresses for IPv4, IPX and OSInet this amounts to 15%.
When you then take into account that there is work being done on dynamic assignment of address ranges by need and anti-fragtional measures. The endresult is hopefully that we will never need to give IP addresses any more thoughts than we today gives the actual routing of packets.:-) = I am happy
:^) = I am happy with my big nose
C:\> = I am happy with my OS
Well, there are already several 6-to-4 and 4-to-6 gateway sites. This one is one example. If there was a site that was only accessable through IPv6 you could use a service like that to access it over any IPv4 host.
Also, if you have an IPv6-capable host you can use a tunnel broker (such as Hurricane Electric's free service) to achieve connectivity to IPv6 sites over IPv4.
So you really don't need an IPv6-capable ISP to access IPv6 hosts, although it's cleaner that way of course.