Print

[wildlava]

I've been anxious to see improved pings via IPv6 to Google using my Denver HE tunnel, and it appears that it is indeed now improved. I used to get ping times of ~70ms to Google (~20ms on IPv4), and packets were going through the Chicago node after Denver before hitting Google. Now I get:

traceroute to google.com (2607:f8b0:400f:802::2000), 30 hops max, 80 byte packets
 1  xxx.xxxx.net (2001:470:xxxx::1)  0.338 ms  0.409 ms  0.505 ms
 2  xxxx.tunnel.tserv1.den1.ipv6.he.net (2001:470:xx:xxx::1)  20.273 ms  20.356 ms  20.716 ms
 3  ge2-19.core1.den1.he.net (2001:470:0:24b::1)  31.797 ms  31.811 ms  31.796 ms
 4  2605:6c00:303:303::3 (2605:6c00:303:303::3)  21.940 ms  21.952 ms  21.940 ms
 5  2001:4860::1:0:24db (2001:4860::1:0:24db)  46.129 ms  46.140 ms  46.127 ms
 6  2001:4860:0:1::121f (2001:4860:0:1::121f)  44.066 ms  42.681 ms  42.658 ms
 7  2607:f8b0:400f:802::2000 (2607:f8b0:400f:802::2000)  45.997 ms  44.756 ms  44.743 ms

This is great! But notice the ~25ms delay between #4 and #5? That is the final hop from CoreSite in Denver to Google. Given that there are Google nodes a mere millisecond or so from my IPv4 ISP's peering in Denver, this 25ms delay seems weird.

I am thrilled that progress has been made, but I wanted to ask about this one hop in case it is due to some issue that HE or CoreSite can to look at.