on 27-03-2010 18:11
07-08-2020 12:19 - edited 07-08-2020 12:19
I think it's way too early to expect to see any improvements yet. My understanding is that VM are still investigating the issue. My guess is that it may take weeks or even months for a complete fix, assuming that they even decide to fix it.
I have some automated monitoring in place now that will alert me any time my IPv6 throughput exceeds 100 Mbit/s so at least I will know when/if something changes.
on 07-08-2020 12:23
@ChrisJenkins wrote:I think it's way too early to expect to see any improvements yet. My understanding is that VM are still investigating the issue. My guess is that it may take weeks or even months for a complete fix, assuming that they even decide to fix it.
I have some automated monitoring in place now that will alert me any time my IPv6 throughput exceeds 100 Mbit/s so at least I will know when/if something changes.
Yup, Assuming they decide to fix it.
I am guessing they will not fix it unless it's both quick to find and easy to fix.
If the VM and O2 merger happen then I can't help but think that there will be yet another culling of staff, probably the person assigned to find this problem lol.
on 07-08-2020 12:29
08-08-2020 09:22 - edited 08-08-2020 09:24
Here's my situation as of this morning:
IPV4 196 Mb/s
IPV6 72.5 Mb/s
I've only just set the tunnel up so I can't compare it with previous days.
on 08-08-2020 09:28
72.5 is pretty good by current standards,
What tunnel provider and endpoint?
What kind of VM modem/router do you have?
Are you running it in modem mode?
on 08-08-2020 09:38
Interesting we are seeing pockets of increase speeds in some cases, for myself I figured out getting iperf3 to work properly with multi WANs so it's actually testing the right WAN.
I see no change in my case currently.
6in4 IPv6
root@linksys-wrt3200acm:~# iperf3 -6 -c bouygues.iperf.fr -R
Connecting to host bouygues.iperf.fr, port 5201
Reverse mode, remote host bouygues.iperf.fr is sending
[ 5] local 2001:470:xxxx:xx::x port 53684 connected to 2001:860:deff:1000::2 port 5201
[ ID] Interval Transfer Bitrate
[ 5] 0.00-1.00 sec 1.09 MBytes 9.15 Mbits/sec
[ 5] 1.00-2.00 sec 1.64 MBytes 13.8 Mbits/sec
[ 5] 2.00-3.00 sec 1.55 MBytes 13.0 Mbits/sec
[ 5] 3.00-4.00 sec 1.84 MBytes 15.4 Mbits/sec
[ 5] 4.00-5.00 sec 1.64 MBytes 13.7 Mbits/sec
[ 5] 5.00-6.00 sec 1.93 MBytes 16.2 Mbits/sec
[ 5] 6.00-7.00 sec 1.52 MBytes 12.7 Mbits/sec
[ 5] 7.00-8.00 sec 1.83 MBytes 15.4 Mbits/sec
[ 5] 8.00-9.00 sec 1.42 MBytes 12.0 Mbits/sec
[ 5] 9.00-10.00 sec 1.90 MBytes 15.9 Mbits/sec
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval Transfer Bitrate Retr
[ 5] 0.00-10.06 sec 17.3 MBytes 14.5 Mbits/sec 13 sender
[ 5] 0.00-10.00 sec 16.4 MBytes 13.7 Mbits/sec receiver
iperf Done.
Wireguard IPv6
root@linksys-wrt3200acm:~# iperf3 -B fc00:bbbb:bbbb:bb01::1:611c -c bouygues.iperf.fr -R
Connecting to host bouygues.iperf.fr, port 5201
Reverse mode, remote host bouygues.iperf.fr is sending
[ 5] local fc00:bbbb:bbbb:bb01::1:611c port 46811 connected to 2001:860:deff:1000::2 port 5201
[ ID] Interval Transfer Bitrate
[ 5] 0.00-1.00 sec 9.18 MBytes 77.0 Mbits/sec
[ 5] 1.00-2.00 sec 10.4 MBytes 87.2 Mbits/sec
[ 5] 2.00-3.00 sec 10.8 MBytes 90.8 Mbits/sec
[ 5] 3.00-4.00 sec 4.23 MBytes 35.5 Mbits/sec
[ 5] 4.00-5.00 sec 8.52 MBytes 71.5 Mbits/sec
[ 5] 5.00-6.00 sec 9.15 MBytes 76.8 Mbits/sec
[ 5] 6.00-7.00 sec 9.62 MBytes 80.7 Mbits/sec
[ 5] 7.00-8.00 sec 9.84 MBytes 82.5 Mbits/sec
[ 5] 8.00-9.00 sec 9.72 MBytes 81.5 Mbits/sec
[ 5] 9.00-10.00 sec 9.70 MBytes 81.4 Mbits/sec
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval Transfer Bitrate Retr
[ 5] 0.00-10.06 sec 95.6 MBytes 79.8 Mbits/sec 819 sender
[ 5] 0.00-10.00 sec 91.2 MBytes 76.5 Mbits/sec receiver
iperf Done.
Maybe we need to start trying to see if routing is anything to do with it, this is my HE tunnel endpoint and the path.
root@linksys-wrt3200acm:~# traceroute 216.66.80.26 -w 1
traceroute to 216.66.80.26 (216.66.80.26), 30 hops max, 38 byte packets
1 * * *
2 nott-core-2b-xe-913-0.network.virginmedia.net (62.255.229.77) 8.871 ms 9.960 ms 8.980 ms
3 * * *
4 * * *
5 * * *
6 m686-mp2.cvx1-b.lis.dial.ntli.net (62.254.42.174) 19.848 ms 21.557 ms 39.957 ms
7 10ge2-4.core1.lon2.he.net (216.66.80.117) 19.042 ms 24.215 ms 17.824 ms
8 tserv1.lon1.he.net (216.66.80.26) 19.877 ms 20.750 ms 18.058 ms
I guess a lot of us are using Hurricane Electric, I know there is also tunnelbroker.ch, as well.
on 08-08-2020 09:39
I'm in Edinburgh, with a superhub 2 running in modem mode. The router is a Netgear 3700 running Openwrt and the HE tunnel is set up on that, connecting to the London endpoint on 216.66.88.98
08-08-2020 11:08 - edited 08-08-2020 11:10
Okay...
I'm in Bracknell and I have a 500 Mbit/s VM Business connection with a Hitron router running in modem mode. My router is an Apple AirPort Extreme (the final model 802.11ac tower with the latest [final] firmware). Over IPv4, depending on how I measure, I can get up to 550 MBit/s downstream and 35 Mbit/s upstream.
My primary tunnel is Hurricane Electric (tunnelbroker.net) to the original London endpoint (216.66.80.26). Originally my tunnel terminated at the AirPort Extreme router but as part if my recent investigations (and avoidance measures) for this issue I have changed my setup...
I now have *two* separate tunnels active from my home location:
1. My original HE tunnel to the same endpoint but now terminated on my Synology NAS which is acting as a router and firewall for IPv6 traffic *only* (all IPv4 traffic [including the tunnelled protocol 41 traffic of course] still goes via the AirPort Extreme and the Hitron over my VM connection.
Over this tunnel my typical performance (it varies marginally over time) is:
HTTP: 28 Mbit/s
IPERF3 (single stream): 22 Mbit/s
I run automated http speed tests on this connection every 20 minutes and log the results for later analysis.
An IPv4 traceroute from this system to the tunnel endpoint looks like this:
traceroute to 216.66.80.26 (216.66.80.26), 30 hops max, 60 byte packets
1 router.home.thejenkinsfamily.org.uk (10.0.200.1) 0.329 ms 0.358 ms 0.394 ms
2 10.8.28.1 (10.8.28.1) 9.962 ms 13.961 ms 13.957 ms
3 winn-core-2a-xe-830-0.network.virginmedia.net (62.253.122.149) 37.699 ms 37.708 ms 38.934 ms
4 * * *
5 m686-mp2.cvx1-b.lis.dial.ntli.net (62.254.42.174) 21.903 ms 21.917 ms 21.864 ms
6 10ge2-4.core1.lon2.he.net (216.66.80.117) 21.716 ms 20.798 ms 20.763 ms
7 tserv1.lon1.he.net (216.66.80.26) 19.283 ms 12.900 ms 14.504 ms
An IPv6 traceroute to 'speedtest6.tele2.net' (for http tests) looks like this:
admin@Luthien:~$ traceroute -6 speedtest6.tele2.net
traceroute to speedtest6.tele2.net (2a00:800:1010::1), 30 hops max, 80 byte packets
1 tunnel238782.tunnel.tserv5.lon1.ipv6.he.net (2001:470:1f08:2df::1) 16.144 ms 16.099 ms 18.050 ms
2 10ge3-16.core1.lon2.he.net (2001:470:0:67::1) 18.929 ms 18.915 ms 20.154 ms
3 2001:7f8:4::4e9:1 (2001:7f8:4::4e9:1) 20.126 ms 20.137 ms 20.122 ms
4 ams-core-2.bundle-ether5.tele2.net (2a00:800:0:1::7a:1) 25.281 ms * *
5 * * *
...
30 * * *
Not very enlightening!
And the same to 'ping6.online.net' (iperf3 server) looks like this:
admin@Luthien:~$ traceroute -6 ping6.online.net
traceroute to ping6.online.net (2001:bc8:1::40), 30 hops max, 80 byte packets
1 tunnel238782.tunnel.tserv5.lon1.ipv6.he.net (2001:470:1f08:2df::1) 14.185 ms 15.436 ms 16.710 ms
2 10ge3-16.core1.lon2.he.net (2001:470:0:67::1) 18.064 ms 19.295 ms 19.273 ms
3 100ge4-1.core1.lon3.he.net (2001:470:0:3ea::2) 17.906 ms 17.879 ms 17.900 ms
4 e0-52.core1.ams2.he.net (2001:470:0:227::2) 25.977 ms 28.554 ms 29.830 ms
5 * * *
6 2001:bc8:0:1::45 (2001:bc8:0:1::45) 32.326 ms 27.293 ms 26.419 ms
7 2001:bc8:0:1::b2 (2001:bc8:0:1::b2) 31.346 ms 2001:bc8:0:1::b6 (2001:bc8:0:1::b6) 28.701 ms 28.673 ms
8 ping6.online.net (2001:bc8:1::40) 27.130 ms 27.144 ms 27.236 ms
2. A second tunnel to the London endpoint (45.12.68.66) of tunnel broker.ch. This terminates, at my end, at my second Synology NAS and provides IPv6 services only for that unit (not my entire network). The underlying IPv4 traffic for this also, of course, goes via my AirPort Extreme and the Hitron over my VM connection.
Over this tunnel my typical performance (it varies marginally over time) is:
HTTP: 31 Mbit/s
IPERF3 (single stream): 23 Mbit/s
An IPv4 traceroute from this system to the tunnel endpoint looks like this:
admin@Elrond:~$ traceroute 45.12.68.66
traceroute to 45.12.68.66 (45.12.68.66), 30 hops max, 60 byte packets
1 router.home.thejenkinsfamily.org.uk (10.0.200.1) 1.337 ms 1.348 ms 1.429 ms
2 10.8.28.1 (10.8.28.1) 14.114 ms 14.141 ms 14.095 ms
3 winn-core-2a-xe-831-0.network.virginmedia.net (62.253.122.161) 15.335 ms 15.338 ms 19.324 ms
4 * * *
5 m686-mp2.cvx1-b.lis.dial.ntli.net (62.254.42.174) 46.788 ms 48.046 ms 46.690 ms
6 213.46.174.38 (213.46.174.38) 19.115 ms 19.577 ms 19.388 ms
7 * * *
8 185.245.80.0 (185.245.80.0) 18.196 ms 19.121 ms 19.019 ms
9 45.12.68.254 (45.12.68.254) 21.172 ms 18.852 ms 18.763 ms
10 * * *
...
30 * * *
Not so useful.
An IPv6 traceroute to 'speedtest6.tele2.net' looks like this:
admin@Elrond:~$ traceroute -6 speedtest6.tele2.net
traceroute to speedtest6.tele2.net (2a00:800:1010::1), 30 hops max, 80 byte packets
1 2a0c:3b80:7b04:e3::1 (2a0c:3b80:7b04:e3::1) 16.074 ms 15.939 ms 17.159 ms
2 2a0c:3b80:554b:2d0c::44fe (2a0c:3b80:554b:2d0c::44fe) 17.138 ms 17.078 ms 18.317 ms
3 40ge1-3.core1.lon2.he.net (2001:7f8:4::1b1b:1) 18.272 ms 26.008 ms 27.271 ms
4 2001:7f8:4::4e9:1 (2001:7f8:4::4e9:1) 21.896 ms 20.515 ms 21.771 ms
5 * ams-core-2.bundle-ether5.tele2.net (2a00:800:0:1::7a:1) 26.900 ms *
6 ams-core-2.tengige0-5-0-0.tele2.net (2a00:800:1010:8::1) 28.087 ms !X * *
And the same to 'ping6.online,.net' looks like this:
admin@Elrond:~$ traceroute -6 ping6.online.net
traceroute to ping6.online.net (2001:bc8:1::40), 30 hops max, 80 byte packets
1 2a0c:3b80:7b04:e3::1 (2a0c:3b80:7b04:e3::1) 13.869 ms 14.983 ms 18.866 ms
2 2a0c:3b80:554b:2d0c::44fe (2a0c:3b80:554b:2d0c::44fe) 18.828 ms 20.110 ms 20.077 ms
3 40ge1-3.core1.lon2.he.net (2001:7f8:4::1b1b:1) 20.115 ms 21.267 ms 20.017 ms
4 100ge4-1.core1.lon3.he.net (2001:470:0:3ea::2) 19.959 ms 19.841 ms 19.796 ms
5 e0-52.core1.ams2.he.net (2001:470:0:227::2) 27.621 ms 28.980 ms 31.463 ms
6 * * *
7 2001:bc8:0:1::45 (2001:bc8:0:1::45) 30.632 ms 30.577 ms 2001:bc8:0:1::37 (2001:bc8:0:1::37) 28.977 ms
8 2001:bc8:0:1::b6 (2001:bc8:0:1::b6) 28.962 ms 28.831 ms 32.733 ms
9 ping6.online.net (2001:bc8:1::40) 32.690 ms 32.632 ms 31.268 ms
Continued in following post due to text length restriction of the forum...
on 08-08-2020 11:09
I also have a third tunnel to Hurricane Electric's second London endpoint (216.66.88.98). This tunnel runs between a Linux cloud hosted VM (in the Oracle Cloud London data centre) to the HE endpoint and does not traverse any VM infrastructure. This VM has a 2 Gbit/s bandwidth for Internet connectivity.
Using the same IPv4 measurement techniques as I do from home I typically see around 816 - 970 MBit/s for http and iperf3.
For IPv6 over the tunnel I see between 594 and 619 Mbit/s. Slower than IPv4 for sure but [perfectly acceptable given I am using a tunnel.
The IPv4 traceroute to the tunnel endpoint looks like this:
admin@jfoci1:~$ traceroute 216.66.88.98 9:58:51
traceroute to 216.66.88.98 (216.66.88.98), 30 hops max, 60 byte packets
1 140.91.200.2 (140.91.200.2) 0.154 ms 140.91.200.6 (140.91.200.6) 0.101 ms 140.91.200.63 (140.91.200.63) 0.137 ms
2 217.163.47.106 (217.163.47.106) 1.179 ms 1.195 ms 1.089 ms
3 ae52.edge7.London1.Level3.net (217.163.47.105) 1.714 ms lag-106.ear4.London2.Level3.net (212.187.219.157) 1.016 ms 1.065 ms
4 ae-2-3211.edge4.London1.Level3.net (4.69.141.234) 1.238 ms ae-1-3111.edge4.London1.Level3.net (4.69.141.230) 1.165 ms ae-2-3211.edge4.London1.Level3.net (4.69.141.234) 1.198 ms
5 10ge2-1.core1.lon2.he.net (216.66.88.237) 1.288 ms 1.303 ms 19.959 ms
6 tserv1.lon2.he.net (216.66.88.98) 1.238 ms 1.338 ms 1.264 ms
And the IPv6 routes to the test servers look like this:
admin@jfoci1:~$ traceroute -6 speedtest6.tele2.net 9:59:50
traceroute to speedtest6.tele2.net (2a00:800:1010::1), 30 hops max, 80 byte packets
1 tunnel595509.tunnel.tserv1.lon2.ipv6.he.net (2001:470:1f1c:697::1) 2.924 ms 3.915 ms 4.853 ms
2 10ge3-20.core1.lon2.he.net (2001:470:0:320::1) 18.582 ms 18.926 ms 18.816 ms
3 2001:7f8:4::4e9:1 (2001:7f8:4::4e9:1) 5.408 ms 5.523 ms 5.482 ms
4 * * *
5 * * *
6 * * *
7 * * *
8 ams-core-2.tengige0-5-0-0.tele2.net (2a00:800:1010:8::1) 9.928 ms !X * *
admin@jfoci1:~$ traceroute -6 ping6.online.net 10:00:17
traceroute to ping6.online.net (2001:bc8:1::40), 30 hops max, 80 byte packets
1 tunnel595509.tunnel.tserv1.lon2.ipv6.he.net (2001:470:1f1c:697::1) 3.494 ms 5.016 ms 6.160 ms
2 10ge3-20.core1.lon2.he.net (2001:470:0:320::1) 18.693 ms 18.650 ms 18.616 ms
3 100ge4-1.core1.lon3.he.net (2001:470:0:3ea::2) 7.517 ms 6.993 ms 6.994 ms
4 e0-52.core1.ams2.he.net (2001:470:0:227::2) 18.933 ms 20.727 ms 17.106 ms
5 * * *
6 2001:bc8:0:1::45 (2001:bc8:0:1::45) 20.811 ms 15.863 ms 2001:bc8:0:1::37 (2001:bc8:0:1::37) 16.288 ms
7 2001:bc8:0:1::b2 (2001:bc8:0:1::b2) 16.190 ms 16.669 ms 2001:bc8:0:1::b6 (2001:bc8:0:1::b6) 16.310 ms
8 ping6.online.net (2001:bc8:1::40) 15.549 ms 15.755 ms 25.457 ms
I don't see any obvious issues in the routing but frankly without knowing a lot more detail than we have it is hard to draw any real conclusions other than 'VM network is terrible for 6in4 traffic'.
on 12-08-2020 07:56
Last few times I have tested my speed has been down around 2-3 meg... Far lower than normal lol..
Perhaps they fixed it, and it turns out that the whole thing was contended to begin with so someone put in the limit to give everyone a fair shake. Wouldn't that be a kick in the teeth :P.