Re: Verizon FIOS IPv6?

[Stephen Frost <sfrost () snowman net>]

* Valdis.Kletnieks () vt edu (Valdis.Kletnieks () vt edu) wrote:
> On Thu, 09 Jan 2014 08:41:30 -0500, Stephen Frost said:
> > I'm really curious how *that* is working out.  My IPv6 tunnel is only a
> > ms or two slower than IPv4 (and it's all sub-15ms), but there is
> > something very odd if the tunnel is *faster*.  Have you tried working
> > out where the difference is coming from (eg: mtr or similar)?
>
> May be different routing based on where the tunnel leads?

Sure, entirely possible, but I'd be investigating into why because
clearly there's a better ipv4 route than that being used, if ipv6
tunneled over ipv4 is faster.  A bit of difference is fine, but it
sounded like more than 'a bit'.

> I recently got Comcast residential native IPv6 working (sort of - bricked a
> Belkin reflashing dd-wrt, but plugging the laptop into the cablemodem directly
> does dhcpv6 just fine).  IPv6 to work is about 30% faster than IPv4, because
> cogent routes v6 for us better:

Fun times.  Neat to hear of folks getting native IPv6 tho.

> 4 pos-1-4-0-0-cr01.ashburn.va.ibone.comcast.net (2001:558:0:f7c5::1) [AS7922] 29.282 ms 
> pos-1-2-0-0-cr01.ashburn.va.ibone.comcast.net (2001:558:0:f6cd::1) [AS7922] 28.307 ms 
> pos-1-4-0-0-cr01.ashburn.va.ibone.comcast.net (2001:558:0:f7c5::1) [AS7922] 29.292 ms
> 5 * * be-12-pe04.ashburn.va.ibone.comcast.net (2001:558:0:f534::2) [AS7922] 26.520 ms
> 6 2001:559::16a (2001:559::16a) [AS7922] 61.890 ms 52.396 ms 2001:559::176 (2001:559::176) [AS7922] 60.563 ms
> 7 2001:550:2:2f::a (2001:550:2:2f::a) [AS174] 58.846 ms 54.715 ms 54.692 ms
> 8 isb-border.xe-5-0-0.155.cns.ipv6.vt.edu (2607:b400:f0:20::5) [AS1312] 46.999 ms 44.160 ms 43.343 ms
> ..
>
> 4 68.86.94.29 (68.86.94.29) [AS7922] 35.819 ms 68.86.91.149 (68.86.91.149) [AS7922] 34.560 ms *
> 5 * * *
> 6 te0-0-0-19.ccr21.atl02.atlas.cogentco.com (154.54.10.233) [AS174] 81.795 ms 77.678 ms 82.887 ms
> 7 be2053.mpd22.atl01.atlas.cogentco.com (154.54.3.145) [AS174] 83.760 ms be2051.ccr22.atl01.atlas.cogentco.com 
> (154.54.0.161) [AS174] 88.056 ms 87.991 ms
> 8 be2169.ccr22.dca01.atlas.cogentco.com (154.54.31.98) [AS174] 89.785 ms 102.779 ms 106.434 ms
> 9 be2177.ccr41.iad02.atlas.cogentco.com (154.54.41.205) [AS174] 85.708 ms be2113.ccr41.iad02.atlas.cogentco.com 
> (154.54.6.169) [AS174] 82.884 ms be2176.ccr41.iad02.atlas.cogentco.com (154.54.41.53) [AS174] 82.888 ms
> 10 38.127.193.146 (38.127.193.146) [AS174] 87.885 ms 91.767 ms 96.178 ms
> 11 isb-border.xe-5-0-0.155.cns.vt.edu (192.70.187.149) [AS1312] 98.993 ms 88.320 ms 89.401 ms
>
> Apparently, going Blacksburg-ashburn-blacksburg is a lot faster than
> blacksburg-ashburn-atlanta-DC-blacbksburg. Who'd thunk it? :)

68.86.94.29 appears to be Georgia, not Ashburn..?- so where is it going
to Ashburn and, probably more importantly, why?  That's more than a
little bit out of your way to go from Blacksburg to Blacksburg...

Does most of your IPv4 traffic flow through Georgia?  Seems like that
might be the real issue here, going
Blacksburg-Atlanta-DC-Ashburn-Blacksburg

Curiously, doing my own tests from here (not far outside Ashburn):

IPv4 to VT: ~21ms
IPv4 to Ashburn IPv6 gateway: ~7ms
IPv6 to VT: ~18ms

All going through DC to Ashburn, unsurprisingly.  It's slightly faster,
as it turns out, but I'm not going to quibble over 3ms.

Have to admit, not sure I could stand 100ms latency between work & home. 

        Thanks,

                Stephen

Attachment: signature.asc
Description: Digital signature