My theory pertaining to the outline of PSN Sony Released
diagram posted here among other places:
Wouldn’t this be problematic with the then current P2P nature of psn? For instance, on GT5 up to 16 users can be connected at once. This game also has voice chat. From what I can tell there are 4-5 contiguous connections per user @ once. PSN also seems to only authenticate connections to PSN once when u sign on, but never to other users thus these p2p connections aren’t authenticated individually, they just count on the fact that if a user is connected to psn, they are permitted to connect to others connected to psn without specifically signing each connection to eachother. In order to sign and secure each connection to each player individually then again to psn would present overhead problems due to authenticating and dedicating ports (using sonys severly flawed iteration of upnp and its tendancy not to close ports properly). Their work-around seems to be to simply sign two ports and stream other players imput to port 80 for non sensitive data. Long story short, I would submit that Sony’s reluctance to update and secure its services was due to the fact that they depended on exploits in the outdated software to allow insecure “signed” connections between users as a means of using users’ bandwidth and computing power instead of their own. This depends on keeping all user data in one huge sandbox effectively depending on one barrier to access. I would also submit that the “application layer” in the above diagram was effectively to the side of the web server and not behind a separate layer of firewall whereas the diagram above implies that its behind two firewalls. The conclusion being, there are two routes of attack, both of which are compromised if the data serer is accessed because of the fact that it does no ever re-authenticate. Am I completely off base here?