This post is just so the community knows about what I have been working on. I consider the issue both interesting and closed.
For about three weeks we have worked with Cisco engineers to address call dropouts for our VPN phones. Calls typically dropped out between 30 seconds to 5 minutes in and this happened in over 80% of cases. We were able to see that internal VPN calls (calls from our VPN phones to other phones in different offices connected to our UC540) only had the issue less than 25% of the time. Calls to external numbers from non-VPN phones rarely failed if ever.
The engineer identified a CPU spike to 100% during our external VPN calls but even after hours of debugging could not identify the cause of the spike. However, when the spike happened, call quality deteriorated and the call became very unstable and eventually dropped. As the cause could not be identified and there were a few other small issues present in the system, we replaced the UC with an identical unit and tested again. Unfortunately, all issues occured on the replacement system just as they did on the original system. All firmwares were up to date. Everything is and was CCA compliant.
Whilst we couldn't identify the cause of the problem, an upgrade of the SPA525G2 phones solved the problem. We were using 7.4.9a firmware. The issue went away when we upgraded to 7.5.1(RC1). We don't know why but now our VPN calls work and previously they didn't.
