Failing to plan is planning to failIPv6 Considered a Problem by Some Users.
I have a Google Blog Search Alert looking for posts over IPv6 in my RSS reader. What strikes me is the number of posts explaining how to disable IPv6 in Windows Vista, MacOSX, Ubuntu and other flavours of Linux.
It looks like disabling IPv6 makes web browsing faster for a lot of people, independently of which operating system is being used. One guy even wrote in one of his posts "In order to fix this problem". IPv6 was supposed to be a solution, not a problem.
He's right, in a sense. IPv6 is a solution to a very challenging problem, the imminent exhaustion of the IPv4 address pool.
But...
The problem being alluded to in those posts explaining how to disable IPv6 is not that, but a completely different one.
That problem is that all too often, when a new computer running Vista, OS X or some flavour of Linux is plugged into an existing home network, Internet performance is severely impacted.
In turn this problem stems from a number of contributory factors, including:
- The new computer has IPv6 enabled by default.
- The existing home network does not have IPv6 capability.
- And neither does the ISP providing Internet service.
The IPv6 enabled computer, when asked to connect to, say, www.google.com, will first do an AAAA look-up to get the IPv6 address. On an IPv4 only home network, this should fail very quickly with a negative DNS response being returned. This, in turn, causes the client computer to fall back to IPv4 and repeat the DNS look-up, looking for an A record.
What appears to be happening is that the original AAAA look-up does not cause a negative response, but either returns the IPv6 address or simply times out. If the IPv6 address is returned, the computer will attempt to connect to it and will time out before falling back to IPv4. In either case, this adds a significant delay before the fall back to IPv4, hence the network performance will appear to be very slow.
The AAAA look-up may be incorrectly handled by a border NAT router, or by some upstream resolver.
What all of this illustrates is a complete lack of coordination among ISPs and providers of networking equipment.
IPv6 has been known about for years and it is surely not unreasonable to expect ISPs to have IPv6 deployment strategies, or for suppliers of network equipment to have engineered them to degrade gracefully when faced with IPv6 requests on an IPv4 only network.
This lack of planning, which has led to a widespread lack of backward compatibility, is the root cause of the issue here.
It is also, perhaps, an indication of the priority given to IPv6 by suppliers both of network hardware and of network connectivity - none at all.
IPv6 will not succeed if adopted piecemeal in this way, with suppliers of end point devices like computers and games consoles racing ahead, while ISPs and suppliers of network hardware lag behind.
By the time networks and border routers are ready (if they ever are), most users will have found one of those blog posts describing how to fix the problem of IPv6 and will have turned it off.
The real problem with IPv6 is the lack of any widely accepted plan or deployment roadmap.
And, as a former colleague of mine used to opine with irritating frequency, failing to plan is planning to fail.
Category: IPv6
Technorati: IPv4 IPv6
1. Mark Dowling14/08/2008 15:26:46
Homepage: http://cork2toronto.blogspot.com
And what was working with Roy Keane like?
http://news.bbc.co.uk/sport3/worldcup2002/low/team_pages/rep_of_ireland/newsid_2006000/2006026.stm
Unable to post a comment? Please read this for a possible explanation...
-