We are running A site-to site VPN solution between our main facility and our colo. we are using 2 3620 routers with VPN accelerator cards. Everything was fine until yesterday. During the day we noticed a constant degradation with our connection. Ping times were going down as well as server response times until the connection was virtually useless. The connection never actually went completely "down" but it may as well have. I resolved the problem with a simple reboot of one of the routers. Has anyone ever experienced this? It was almost like a memory leak or buffer problem. What kind of Forensics can i run on the routers that may explain?