IPv6 – Giving preference to IPv6 versions of websites

Now that I have my IPv6 Tunnel Broker connection up and running, I next turned to how to see what traffic was going out over that connection. If I used a hostname of ipv6.google.com, it was reasonable to assume that traffic for that URL would go out over the IPv6 connection. Other than that, without using Wireshark, I wasnt able to see how to identify which websites were IPv6 accessible. Some websites I went to would tell you what source IP address that they saw you coming from. If web browsers are supposed to go to IPv6 websites first, by typing www.google.com, you should theoretically end up at the same site as if you were going to ipv6.google.com. That isnt what I found as I starting going to different websites that were supposed to be IPv6 enabled.

Here are some of the ones I found that will show you if you are accessing them via IPv6 or IPv4 –

http://ipv6-test.com/ – tests IPv4 and IPv6 connections / dual stackhttp://test-ipv6.com/ – another IPv4/IPv6 test site
http://www.ipv6security.nl/ – IPv6 info and test for IPv6 connection
http://ipv6-speedtest.net/ – IPv6 Speed test
http://www.wireshark.org/tools/v46status.html – Shows the ability to reach IPv4 and IPv6 versions of the wirehsark website
http://www.ipv6chicken.com/ – IPv6 test that also tests your MT setting
http://www.fnlsix.net/ – FermiLab Test IPv6 test page

That is one way of verifying the traffic but that only helps me on those sites that are displaying the source URL you are detected to be using. Everything I read said that the web browsers were all geared to look for an IPv6 AAAA record first and use that before going with an IPv4 A record. With Firefox, I found several Add-on’s that showed you a 4 or a 6 depending on what type connection the web browser was using to access a given site. I with with 4or6 and SixOrNot. Going with one should have been sufficient but I wanted verification that I was actually using IPv6 whenever possible without having to resort to packet sniffing.

If everything was working right, I would expect that both Add-ons would both indicate 4 or 6 at the same time. About 75% of the time, they would agree but not as much as I would have expected. What I ended up having to do was to plug in the IPv6 address of a DNS server at Hurricane Electric, followed by the IPv4 address of another DNS server at Hurricane Electric into the DNS server configuration for my laptop, overriding what the DHCP server was handing out. Once I did that, both indicates consistently showed 4 or 6 at the same time. Will have to do some additional research as time permits to figure out why this happened. Either the DNS servers I was initially referencing werent setup correctly to forward IPv6 AAAA records, there is a problem with IPv6 Glue on some of the domains I was testing with or I have another problem that I wasnt able to identify at this point.

One of the next things I am going to look at is seeing if there might be “broken” implementations of IPv6 that would explain why I seemed to have inconsistent results on results to an IPv6 AAAA record. I will setup a spreadsheet to track each of the domains that I try along with the DNS configuration on the laptop and the results that I received during that attempt. Will also run Wireshark at each attempt to see what traffic is getting returned. To make things are reproduce-able as possible, I will put down as one of my test steps to clear out the DNS resolution cache to make sure that I am going for AAAA and A DNS records with each attempt from the web browser.

Send to Kindle
This entry was posted in Blog Entries and tagged . Bookmark the permalink.