There has been some discussion whether it would be a good idea to use
the IPv4 mapped IPv6 prefix (::ffff:0:0/96) or the IPv4 compatible
prefix (::/96) as the prefix used for a NAT-PT/NAT64 translator that
lets IPv6 clients talk to IPv4 servers.
In order to see if such an approach would be workable we want to test
the behavior of different operating systems when applications are fed
AAAA records with these addresses in them. We already have some
testers for common OSes (Linux, Win XP/Vista, MacOS 10.5, Symbian) but
we can use more measurements for less common OSes. The IETF network is
conveniently available in dual stack and IPv6-only flavors.
We've set up the following test to determine how operating systems
handle IPv4 mapped IPv6 addresses (::ffff:0:0/96) and IPv4 compatible
addresses (::/96) they encounter in AAAA records.
Please go to http://test.ams.runningipv6.net/
(Make sure you're reaching this page over IPv6 as indicated by your
source address that the server echos back.)
There, you can download an Excel (XLS) file, please fill in your
details and those of the machine you're testing with, and set up your
system with the combination of settings an reachability that you want
to test, as listed in the spreadsheet. Since most people will run in a
default configuration, please test at least the defaults for IPv4 and
IPv6 configured on the system / active interface along with dual stack
and IPv6-only reachability. There are three other configurations that
you can test and you can add more if you like.
Visit the 8 URLs listen on the page with the selected settings and
record the results in the appropriate column.
Repeat for other settings you want to test.
If the system uses normal A or AAAA records (which are available in
different combinations for the test2 - test4 URLs) you'll return to
the test page, if the system uses a mapped (or compatible) address
successfully you'll see the IETF or IANA pages. However, the most
interesting cases are those where there is a timeout on an IPv6
network, in that case the host may be trying to send IPv6 packets with
a destination in the /96 prefix in question.
You can do this at any time, but preferably before Monday so we can
compile and send out the results before NAT64-related discussions
commence.
Use the nameserver address in the XLS file if you're experiencing DNS
issues, but this shouldn't be necessary.
We would also like to test with at least a few OSes in such a way that
we can determine if the OS generates any IPv6 packets with compatible
or mapped destination addresses. To do this, please find me (in
addition to email you can reach me on Jabber and other instant
messaging networks as [EMAIL PROTECTED]) and we'll test with a direct
ethernet cable between the system to be tested and my laptop, which
will act as an IPv6 router and intercept all packets sent over the
ethernet port. This is what the second half of the XLS file is for.
_______________________________________________
Int-area mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/int-area
