Originally Posted by
GUWonder
That mention about consumer ISPs who have significant CGNAT on their IPv4 is an interesting one. Would you say that could be a part of why some of these FT-specific Cloudflare connection blocks — even of non-VPNs — may be geographically clustered with their impact on FTers and others trying to connect with the site?
It's certain possible. IP reputation tools can down score the address used by an IPv4 CGNAT service where they see patterns of dubious activity just in the same way that a commercial VPN (which will also certainly be CGNAT). The one difference is that the connections are less likely to be so anonymous.
My own personal assessment of the typical averages of the IP reputation of categories of each type of service.
Anonymous - the extent to which users are trying to hide their ID
Malicious - the extent of usage that is malicious or naughty
Longevity - how long users would typically use such a service
Geohopping - the utility of the service to permit a user to jump elsewhere globally
Commercial VPN:
Anonymous high, Malicious high, longevity high, geohopping high
Public WiFi:
Anonymous medium, Malicious medium, longevity low, geohopping n/a
ISP CGNAT:
Anonymous medium, Malicious medium, longevity high, geohopping n/a
ISP individual user IP:
Anonymous low, Malicious low, longevity high, geohopping n/a
I am sure we can all argue cases for all of the above to be different low/medium/high but I am trying to set out typical values.
Many ISPs who are using CGNAT for IPv4 addresses would hopefully be running IPv4/IPv6 dual stack in which case there's the additional dimension that v6 connections could individual and personal whereas v4 could be CGNAT.