Subj : IPv6 day To : Bj”rn Felten From : Michiel van der Vlist Date : Wed Jun 08 2011 15:31:22 Hello Bj”rn, On Wednesday June 08 2011 14:56, you wrote to me: MvdV>> Halfway into the ride and nothing much seems to be happening. MvdV>> No reports about massive failures, etc. BF> No big surprise IMHO. I was under the impression that a lot of BF> major sites would go IPv6-only for 24 hours. No, the objective of the test was not to test an IPv6 only situation. The objective was to test dual stack implementations. BF> But all they did was launch IPv6. Big deal. NOT! Well, actually it is. Or so it was thought. BF> Surely they could have launched their IPv6 pages anyhow and any time? They could, but they were afraid to lose customers. It is estimated that 0.5% of all clients run on systems that think they have IPv6 but in reality do not, or have a broken implementation. I had this happen in the early phase of experimenting with tunnels. My IPv6 was broken and what happens then when attempting to access a dual stack content service is not something you like. The client attempts to connect through IPv6 and the connection times out after 30 seconds or so. Then it tries IPv4. If it is a single file that has to be fetched, this is just bearable. But if it is a website with many items to load, then the 30 second time out will occur for EACH item that is to be loaded. The delay will add up to 5 minutes for a site with 10 pictures. That is horrible. Customers will walk away from a site that loads slowly and go to the competetion who does not have IPv6 enabled. The estimate is 0.5%. But no one realy knows. The idea of IPv6 day is to smoke out those clients with faulty IPv6 implementations, while giving every content provider the same handicap. That *I* have no problems is no surprise. I have delved into the subject and thouroughly tested my IPv6 connections. But how about auntie Margie and uncle George, who just bought a shiny new computer with Windows 7 (IPv6 enabled by default) who are completely illerate about what is under the hood? Their computer may attempt to connect via a shakey 6to4 relay and experience strange delays or timeouts.... Cheers, Michiel --- GoldED+/W32-MINGW 1.1.5-b20070503 * Origin: 2001:470:1f15:1117::1 (2:280/5555) .