The function of media bypass is to allow endpoints to send media traffic (audio) directly to the PSTN gateway, and use the mediation pool only for signaling. This greatly improves the number of calls that a mediation server can handle. When media bypass is not selected, each co-located mediation server can handle up to 155 simultaneous calls (a dedicated mediation server can handle up to 1,500 calls). With media bypass enabled, the performance is greatly improved.

Media bypass and CAC do not go hand in hand, and this is why the always bypass option is disabled when Enable CAC is selected. It is said that CAC and media bypass are mutually exclusive. There are scenarios where we might possibly not need CAC in a deployment. It is worth going through the options and looking at a few scenarios.

When would it be possible to use media bypass in a deployment where QoS and CAC are implemented?

In a large deployment with several sites and PSTN trunks, or simply in a huge campus with a PSTN trunk, it could be of real benefit to have those clients with the same bypass ID as the trunk to bypass the mediation server. This will offload the mediation servers and keep resources available for those calls that come across the WAN, which need CAC to be maintained.

Whenever QoS and CAC are implemented together, we need to use the Use Site and Region Information option where there has to be a match between the caller and the called party's bypass ID.