the MC maintains the station database, which includes information such as the current point of attachment, station's credentials, IP address, as well as the mobility protocol state associated with the station. These states are dynamically updated based on the authentication, address assignment, or mobility signaling events. Redundancy is required to handle the case when the MC fails or is taken down for service, thus eliminating the single point of failure phenomena. This level of reliability provides a switch-over mechanism (i.e. backup MC taking over operation) that does not require the mobile station, switch, or Mobility Oracle to be aware or take any additional action.
MC redundancy uses an active/passive approach and provides a scalable high availability with instantaneous fail over support. When the active MC fails, the backup MC takes over immediately to maintain the states for the stations. The MC sets up the tunneling operation on the MTE, which is responsible for the encapsulation and decapsulation of packets to/from the switch or another MTE for the stations in the sub-domain.
During normal operation, the active MTE does the encapsulation/decapsulation. Since all of the station states are completely replicated on the backup MC, the forwarding path on the backup MTE is already set up to handle packets for the station. This means there is minimal packet loss, which happens during the period between active MC going down and the backup MC detecting the condition and taking over. The number of stations on the MC does not affect the performance of switchover, which is based on the detection latency only. Normally, the active MTE advertises the station if the point of presence is at the MTE. When the backup MTE takes over, it will need to start advertising the station.
There are two components of MC redundancy:
- Unresponsiveness Detection – This is a mechanism to detect when a active MC is down.
- State Synchronization – This mechanism is used transfer the stations' states between the MCs
State synchronization is accomplished by the transfer of the station states from the active MC to the backup MC. There are two types of operations: station update and bulk synchronization. When a station state is created, changed, or deleted, the station's states, or only the status code/delta, are transferred. This update messaging happens when both MCs are in operation. In the case when a backup MC initially comes up, bulk synchronization happens. The backup MC sends a request to the active MC, and downloads the entire database from the active MC. The states of all the stations are reliably transferred using SCTP or application-based reliability mechanism.