Facility MEPs have been introduced to improve scalability, reduce operational overhead, and provide fate sharing without requiring service MEPs. This allows for fault notification for Epipe services that share a common transport. Facility MEPs recognize failure based solely on ETH-CFM detection mechanisms.
There are a total of four facility MEPs, as described below:
In general, a Facility MEP detects failure conditions using ETH-CFM at the Ethernet Transport layer. The detection is based solely on the MEP entering a fault state as a result of ETH-CC. Conditions outside the scope of ETH-CFM do not directly influence the state of the MEP. However, these outside influences have indirect influence. For example, upon a failure of a port, CCM messages cannot reach the destination. This condition causes the MEP to enter a fault state after the 3.5*interval expires, with the only exception being the acceptance of AIS on a Tunnel MEP. AIS received on all other facilities MEPs are discarded silently when normal level matching targets the local facility MEP.
Facility MEPs are supported as part of a down MEP only. Facility MEPs validate the point to point Ethernet transport between two end points. Facility MEPs do not validate switching functions that are not part of the point to point Ethernet transport. Instead, service MEPs validate switching functions that are not part of the point to point Ethernet transport.
A facility MEP allows for the scaling improvements using fate sharing and leveraging OAM mapping. The OAM mapping functions are part of the fault propagation functions and allow ETH-CFM to move from alarms only to network actions. Service based MEPs are not required to generate AIS in reaction to a facility MEP fault. OAM mapping and generation of fault via fault-propagation means or the AIS function are only available for Epipe services. There is no equivalent AIS generation as part of the facility fault for VPLS, IES, and VPRN. There is no service MEP required to have the SAP transition in the VPLS, IES, and VPRN service context. Normal SAP transition functions does not occur when these services are configured to accept the tunnel fault, or in reaction to a facility fault, where the underlying port or LAG transitions the SAP.
The implementation of facility MEPs must adhere to all platform-specific specifications. For example, sub-second enabled CCM MEPs are supported on port based MEPs. However, any platform restrictions preventing the sub-second enabled MEPs override this capability and require the operator to configure CCM intervals that are supported for that specific platform.
Facility MEPs are created in the same manner as service MEPs, both related to the ETH-CFM domain and association. However, the association used to build the facility MEP does not include a bridge-identifier. The CLI ensures that a bridge ID is not configured when the association is applied to a facility MEP.
Service MEPs and Facility MEPs may communicate with each other, as long as all the matching criteria are met. Because facility MEPs use the standard ETH-CFM packets, there is nothing contained in the packet that would identify an ETH-CFM packet as a facility MEP or Service MEP.
Facility MEPs are not supported on ports that are configured with Eth-Tunnels (G.8031) and only facility MEPs of 1 second and above are supported on the ports that are involved in an Eth-Ring (G.8032).