Re: auto6to4
> 6in4 provide e.g. by tunnelbroker.net or sixxs is a much better ...
Well I'm just trying to make it easy to configure 6to4, not to force
people to use it.
That said,
(a) Those require the user to *really* know what they're doing: to
know about the right web sites, go to one of them request a tunnel
which in some cases requires sacrificing an unblemished lamb, and in
many cases putting the information into the right configuration files,
which in turn requires knowing what those files are and what they look
like and how to tweak and test them, etc. Even AICCU requires
answering some questions at configuration time.
(b) The relative advantages of the two technologies depends upon what
the ISPs are doing. The article you refer to is basically exploring
how the ISPs often screw up 6to4. If ISPs do "the right thing" then
6to4 should be fast and reliable. For my particular situation, I
often find that Teredo and AICCU sixxs tunnels are slower and less
reliable than 6to4, presumably because the involved ISPs are top
notch.
(There is an example below showing mixed results, but I'm
topologically very close to sixxs right now; when I travel, AICCU
performance usually drops through the floor.)
All this script tries to do is make it easy to set up 6to4, not to
suggest that people use it in preference to trickier things that require
reading books and such.
But this does bring up some related questions.
First:
It appears that right now, if hosts A and B each have both IPv4 and IPv6
addresses, with B's both listed in DNS, then by default Debian will try
IPv6 before IPv4 when A tries to connect to B. Would there be a way for
a package like auto6to4 to set some global configuration option so that
programs will try IPv4 before IPv6-via-6to4 when both are available?
Second:
Would it make sense to factor out the automatic radvd configuration
stuff, so that it could be used with things other than 6to4? After all
there is nothing really specific to 6to4 about that part of the script:
all it needs is appropriate IPv6 subnet(s).
Third:
Would it be possible to make a similar script that *fully* automates
obtaining and setting up one of those fancier tunnels, the way the
miredo package completely automates setting up miredo? If that's hard
right now, could it be facilitated by actions on the part of sixxs?
--Barak.
################ Teredo ################
$ sudo auto6to4 stop
$ sudo /etc/init.d/miredo restart
[ ok ] Stopping Teredo IPv6 tunneling daemon: miredo.
[ ok ] Starting Teredo IPv6 tunneling daemon: miredo.
$ oping -6 -c10 ftp.ie.debian.org
PING debian.heanet.ie (2001:770:18:aa40::c101:c141) 56 bytes of data.
56 bytes from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=1 ttl=56 time=254.12 ms
56 bytes from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=2 ttl=56 time=76.63 ms
echo reply from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=3 timeout
56 bytes from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=4 ttl=56 time=76.78 ms
56 bytes from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=5 ttl=56 time=78.10 ms
echo reply from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=6 timeout
56 bytes from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=7 ttl=56 time=76.55 ms
56 bytes from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=8 ttl=56 time=84.24 ms
56 bytes from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=9 ttl=56 time=80.53 ms
echo reply from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=10 timeout
--- debian.heanet.ie ping statistics ---
10 packets transmitted, 7 received, 30.00% packet loss, time 726.9ms
rtt min/avg/max/sdev = 76.551/103.849/254.119/66.322 ms
$ oping -6 -c10 ipv6.google.com
PING dy-in-x63.1e100.net (2a00:1450:400b:c00::63) 56 bytes of data.
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=1 ttl=56 time=204.72 ms
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=2 ttl=56 time=57.37 ms
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=3 ttl=56 time=56.41 ms
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=4 ttl=56 time=60.30 ms
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=5 ttl=56 time=56.85 ms
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=6 ttl=56 time=59.56 ms
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=7 ttl=56 time=59.28 ms
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=8 ttl=56 time=59.32 ms
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=9 ttl=56 time=58.84 ms
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=10 ttl=56 time=58.93 ms
--- dy-in-x63.1e100.net ping statistics ---
10 packets transmitted, 10 received, 0.00% packet loss, time 731.6ms
rtt min/avg/max/sdev = 56.412/73.158/204.723/46.244 ms
################ 6to4 ################
$ sudo /etc/init.d/miredo stop
[ ok ] Stopping Teredo IPv6 tunneling daemon: miredo.
$ sudo auto6to4 start
$ oping -6 -c10 ftp.ie.debian.org
PING debian.heanet.ie (2001:770:18:aa40::c101:c141) 56 bytes of data.
56 bytes from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=1 ttl=61 time=33.36 ms
56 bytes from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=2 ttl=61 time=32.43 ms
56 bytes from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=3 ttl=61 time=32.42 ms
56 bytes from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=4 ttl=61 time=32.43 ms
56 bytes from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=5 ttl=61 time=32.63 ms
56 bytes from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=6 ttl=61 time=32.64 ms
56 bytes from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=7 ttl=61 time=32.42 ms
56 bytes from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=8 ttl=61 time=33.14 ms
56 bytes from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=9 ttl=61 time=36.34 ms
56 bytes from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=10 ttl=61 time=32.90 ms
--- debian.heanet.ie ping statistics ---
10 packets transmitted, 10 received, 0.00% packet loss, time 330.7ms
rtt min/avg/max/sdev = 32.422/33.071/36.336/1.194 ms
$ oping -6 -c10 ipv6.google.com
PING dy-in-x63.1e100.net (2a00:1450:400b:c00::63) 56 bytes of data.
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=1 ttl=56 time=46.72 ms
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=2 ttl=56 time=46.70 ms
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=3 ttl=56 time=46.50 ms
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=4 ttl=56 time=47.95 ms
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=5 ttl=56 time=46.20 ms
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=6 ttl=56 time=45.91 ms
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=7 ttl=56 time=46.68 ms
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=8 ttl=56 time=46.41 ms
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=9 ttl=56 time=46.39 ms
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=10 ttl=56 time=46.14 ms
--- dy-in-x63.1e100.net ping statistics ---
10 packets transmitted, 10 received, 0.00% packet loss, time 465.6ms
rtt min/avg/max/sdev = 45.909/46.560/47.949/0.555 ms
################ AICCU ################
$ sudo auto6to4 stop
$ sudo /etc/init.d/aiccu restart
[ ok ] Restarting SixXS Automatic IPv6 Connectivity Client Utility: aiccu.
$ oping -6 -c10 ftp.ie.debian.org
PING debian.heanet.ie (2001:770:18:aa40::c101:c141) 56 bytes of data.
56 bytes from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=1 ttl=62 time=34.82 ms
56 bytes from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=2 ttl=62 time=40.24 ms
56 bytes from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=3 ttl=62 time=40.68 ms
56 bytes from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=4 ttl=62 time=32.97 ms
56 bytes from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=5 ttl=62 time=36.15 ms
56 bytes from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=6 ttl=62 time=33.48 ms
56 bytes from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=7 ttl=62 time=33.23 ms
56 bytes from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=8 ttl=62 time=33.01 ms
56 bytes from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=9 ttl=62 time=33.72 ms
56 bytes from debian.heanet.ie (2001:770:18:aa40::c101:c141): icmp_seq=10 ttl=62 time=33.47 ms
--- debian.heanet.ie ping statistics ---
10 packets transmitted, 10 received, 0.00% packet loss, time 351.8ms
rtt min/avg/max/sdev = 32.972/35.177/40.682/2.949 ms
$ oping -6 -c10 ipv6.google.com
PING dy-in-x63.1e100.net (2a00:1450:400b:c00::63) 56 bytes of data.
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=1 ttl=57 time=34.22 ms
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=2 ttl=57 time=34.27 ms
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=3 ttl=57 time=33.99 ms
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=4 ttl=57 time=34.74 ms
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=5 ttl=57 time=34.26 ms
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=6 ttl=57 time=36.21 ms
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=7 ttl=57 time=34.01 ms
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=8 ttl=57 time=34.73 ms
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=9 ttl=57 time=34.48 ms
56 bytes from dy-in-x63.1e100.net (2a00:1450:400b:c00::63): icmp_seq=10 ttl=57 time=34.46 ms
--- dy-in-x63.1e100.net ping statistics ---
10 packets transmitted, 10 received, 0.00% packet loss, time 345.4ms
rtt min/avg/max/sdev = 33.994/34.538/36.210/0.642 ms
Reply to: