While the ME concept is very valid from an OAM architectural perspective, the definition of an object class to describe it seems useless and potentially incorrect
The need to have an MaintenanceEntity is not fully clear: a MaintenanceJob can be created between two MEPs to measure the performances of the ME between these two MEPs
Using the MaintenanceEntity would make it difficult to model the multicast OAM tools where e.g., DM is used with multicast DA to measure the delay between one MEP and all its peer MEPs within the same MEG
Moreover, the MEs can be known only by the entity which know all the MEPs within an MEG since it is a relationship between two MEPs in the same MEG
As discussed in previous meetings, the following issues exist if a MaintenanceEntity object class is defined:
MIPs can belong to one or more MEs depending on the "route" of the MEs
in multi-domain scenarios, it is possible to configure in one domain an MEG between "edge" MIPs (and no MEPs): in this case there is not enough information to understand how many MEs exist between these two MIPs
The last two scenarios have been discussed in previous ONF meetings and some drawings done at the whiteboard: I am preparing a couple of slides to describe these scenarios
While the ME concept is very valid from an OAM architectural perspective, the definition of an object class to describe it seems useless and potentially incorrect
The need to have an MaintenanceEntity is not fully clear: a MaintenanceJob can be created between two MEPs to measure the performances of the ME between these two MEPs
Using the MaintenanceEntity would make it difficult to model the multicast OAM tools where e.g., DM is used with multicast DA to measure the delay between one MEP and all its peer MEPs within the same MEG
Moreover, the MEs can be known only by the entity which know all the MEPs within an MEG since it is a relationship between two MEPs in the same MEG
As discussed in previous meetings, the following issues exist if a MaintenanceEntity object class is defined: