When the media servers collocated with major branch offices are locally redundant and the smaller branch offices would use a backup media server located at the Lotus Sametime Unified Telephony cluster nodes then this is called N + k redundancy. N + k media server survivability can be achieved for both, geographically collocated and separated Lotus Sametime Unified Telephony Clusters. In normal operation, the media servers at all branches are operational and handle all the media server traffic from the Lotus Sametime Unified Telephony. Load balancing is achieved by assigning each branch office its own rate area.
In N+k redundancy, N represents the number of primary media servers just as it does in N + 1 redundancy. K represents the number of backup media servers for a given primary media server. This number may be different for each primary media server.
In case of failure, the following takes place depending on the failure type:
Node failure: When a Lotus Sametime Unified Telephony node fails, the other node takes over. The MGCP signaling manager on the partner node still uses media servers in the same manner as before the failure.
Media server failure: When any of the three active media servers fails, the Lotus Sametime Unified Telephony detects this event. If the branch has its own backup, the Lotus Sametime Unified Telephony starts using that media server. If the branch does not have its own backup, or if the backup server also fails, the Lotus Sametime Unified Telephony starts using the centralized backup media server.
Branch failure: When an entire branch fails, the other branches continue to use their own media servers and are unaffected by the failure.
WAN failure: When the WAN connection with a particular branch fails, the other branches are unaffected by the failure.
In both figures, the following elements are present:
Each branch office has its own media server, for a total of three primary media servers.
One centralized backup media server is located at one of the Lotus Sametime Unified Telephony cluster nodes.
Two branches (A and C) also have their own backup media servers. These two branches might represent larger offices, whereas Branch B might be a smaller branch that does not require local redundancy.
As a result, branches A and C have two backup media servers; branch B has one.
Parent topic: Survivability Overview
Media Server Survivability with 1+1 Redundancy