[ntp:questions] NTP over IPv6 with Teredo

Danny Mayer mayer at ntp.isc.org
Mon Feb 18 02:59:58 UTC 2008


Ronan Flood wrote:
> At my workplace we're slipping in IPv6 on our NTP servers, and I've
> been watching the traffic and checking out the clients which pop up,
> to see if they hit any problems when unintentionally using IPv6
> instead of IPv4.
> 
> Native-IPv6 and 6to4-tunnel clients seem OK, but I've found one client
> using a Teredo tunnel (RFC 4380), and it looks like that is something
> to be avoided:
> 
>      remote       refid  st t when poll reach   delay   offset  jitter
> ---------------------------------------------------------------------
> +2001:630:1:101: .GPS.    1 u  621 1024  177  350.111  115.061   2.000
> -2001:630:1:102: .MSF.    1 u  617 1024  177  355.528  114.905   0.734
> -2001:630:1:103: .PPS.    1 u  631 1024  177  575.602  222.278   0.056
> +195.66.241.3    .PPS.    1 u  524 1024  377   22.724  -15.799   0.788
> *195.66.241.10   .IRIG.   1 u  753 1024  377   27.490  -15.864   0.775
>  127.127.1.0     .LOCL.  13 l   28   64  377    0.000    0.000   0.002
> 
> Being a UK client, using a Teredo server on the west coast of the US
> probably doesn't help, but it could also be inherent latency in Teredo.
> 

You might ask why they have configured their NTP server to go to the 
west coast of the US in the first place, irrespective of Teredo since 
network wise it a much longer distance than say some other location in 
London.

All tunnels will have additional delays being Teredo, VPN or something else.

> "server -4" would be wise here :-/
>

Only if the really need to go distances like this.

Danny




More information about the questions mailing list