Насчет циско и MGCP, пишут следующее:
"The MGCP gateway maintains a remote connection to a centralized CUCM cluster, as shown in figure by sending MGCP keepalive messages to the CUCM server at 15 second intervals.
If the active CUCM server fails to acknowledge receipt of the keepalive message within 30 seconds, the gateway attempts to switch over to the next available CUCM server.
If none of the CUCM server responds, the gateway switches into fallback mode and reverts to the default H.323 or SIP session application for basic call control."
Т.е. переключение на резервный сервер занимает минимум 30 секунд после падения канала.
"The MGCP gateway maintains a remote connection to a centralized CUCM cluster, as shown in figure by sending MGCP keepalive messages to the CUCM server at 15 second intervals.
If the active CUCM server fails to acknowledge receipt of the keepalive message within 30 seconds, the gateway attempts to switch over to the next available CUCM server.
If none of the CUCM server responds, the gateway switches into fallback mode and reverts to the default H.323 or SIP session application for basic call control."
Т.е. переключение на резервный сервер занимает минимум 30 секунд после падения канала.