In an effort to overcome the impending problem of IPv4 addresses running out, many ISPs have started to look into using Carrier-Grade Network Translation (CGNAT). Unlike sharing a single IP address at home using a router, CGNAT is implemented at the ISP level. This enables to the ISP to share one IP address among many customers, rather than the current model where each household (ISP account) is assigned their own unique IP address.
If an IP address is shared, then identifying which user of that IP address is responsible to any behavior logged to that address is difficult. This has led to speculation that such a system could be used to obfuscate the behavior of file sharers, as it might be impossible for Intellectual Property enforcers to be certain of who was guilty. At the very least, it might work as an extra layer of security for copyright infringers.
British ISP BT has now attempted to scotch this idea, claiming that “the technology does still allow individual customers to be identified if they are sharing the same IP address, as long as the port the customer is using is also known.”
Going into further detail, a BT spokesman explained that as long as they know the IP, port used and timestamp of the alleged violation, they could identify the unique subscriber involved.
Using a VPN will of course completely hide all such information from an ISP. Interestingly, some VPN providers such as Private Internet Access also use shared IPs, although in this case the purpose is to deliberately make identifying an individual difficult. We suppose that in theory individuals could be identified using the detection methods outlined by BT, but as any ISP- worth its salt (such as PIA) keeps no logs anyway, it should indeed provide an extra layer of protection for users.