What's up with BGP communities?

[Ronan Pigott via NANOG <nanog () lists nanog org>]

Hi,

First off, apologies if I'm completely lost. I'm not a network professional of any kind, just a random with a question 
I can't find the answer to on the internet.

Basically, I was poking around on this website called bgp.tools and many ASN have a page describing BGP communities 
known by the network, e.g. https://bgp.tools/communities/13335.

It includes descriptions for each one:

> 13335:10106 PoP: bos01
> 13335:10358 PoP: cwb03
> 13335:10712 PoP: den04
> 13335:10766 PoP: maa05
> 13335:10920 PoP: zrh02
> [...]

I want to know, where does this supplementary information about the communities come from, in this case the "PoP: blah" 
bits? I don't think it is communicated by BGP directly, so is there some standard out of band mechanism to describe the 
communities? Or am I just ignorant of this BGP feature?

The background is that I live in AZ, USA, and am I subscribed to Cox for my internet service. I noticed that my ping to 
"example.com" (literally) is slower on IPv6 compared to IPv4:

  $ ping -4 example.com
  PING example.com (104.18.27.120) 56(84) bytes of data.
  64 bytes from 104.18.27.120: icmp_seq=1 ttl=251 time=9.58 ms
  64 bytes from 104.18.27.120: icmp_seq=2 ttl=251 time=5.94 ms
  64 bytes from 104.18.27.120: icmp_seq=3 ttl=251 time=7.76 ms
  64 bytes from 104.18.27.120: icmp_seq=4 ttl=251 time=6.85 ms
  [...]
   $ ping -6 example.com
  PING example.com (2606:4700::6812:1b78) 56 data bytes
  64 bytes from 2606:4700::6812:1b78: icmp_seq=1 ttl=57 time=20.0 ms
  64 bytes from 2606:4700::6812:1b78: icmp_seq=2 ttl=57 time=19.0 ms
  64 bytes from 2606:4700::6812:1b78: icmp_seq=3 ttl=57 time=20.8 ms
  64 bytes from 2606:4700::6812:1b78: icmp_seq=4 ttl=57 time=19.8 ms
  [...]
  $ traceroute -IAen4 -z1 example.com
  traceroute to example.com (104.18.26.120), 30 hops max, 60 byte packets
   1  10.3.0.254 [*]  0.249 ms  0.361 ms  0.362 ms
   2  10.40.160.1 [*]  8.978 ms  8.658 ms  7.745 ms
   3  100.127.73.232 [*]  6.957 ms  8.557 ms  6.881 ms
   4  68.1.0.187 [AS22773]  8.067 ms  10.434 ms *
   5  184.183.131.9 [AS22773]  13.818 ms  37.728 ms  17.832 ms
   6  162.158.140.23 [AS13335]  16.822 ms  8.761 ms  16.768 ms
   7  104.18.26.120 [AS13335]  8.813 ms  7.852 ms  9.451 ms
  $ traceroute -IAen6 -z1 example.com
  traceroute to example.com (2606:4700::6812:1b78), 30 hops max, 80 byte packets
   1  2600:8800:1780:c000::1 [AS22773]  0.315 ms  0.360 ms  0.362 ms
   2  2600:8800:17ff:ffff::1111 [AS22773]  6.601 ms  24.911 ms  9.655 ms
   3  2001:578:800:6:8100::818 [AS22773]  31.164 ms  8.727 ms  9.075 ms
   4  2001:578:900:4::2 [AS22773]  8.763 ms  7.661 ms  6.873 ms
   5  2001:578:1:0:172:17:249:32 [AS22773]  29.991 ms  35.838 ms  19.517 ms
   6  2001:578:20:a000::7:1 [AS22773]  21.781 ms  20.676 ms *
   7  2400:cb00:12:3:: [AS13335]  20.474 ms  20.099 ms  19.824 ms
   8  2606:4700::6812:1b78 [AS13335]  20.078 ms  19.747 ms  20.788 ms

It's not really a problem, but still I just want to know... why? I thought it would be kind of odd to have example.com 
hosted only on IPv4 in the Phoenix area so I was trying to see if I could determine whether or not that was the case 
looking at https://bgp.tools/communities/13335?show-prefixes=13335%3a10066, but I think there just simply isn't enough 
information available online for me to find out.

Thanks,

Ronan
_______________________________________________
NANOG mailing list 
https://lists.nanog.org/archives/list/nanog () lists nanog org/message/IC3W2ONG6J7WUKN5QDE6JIOMDNHSCKRB/