NFD

Currently a nexthop entry is as NextHop(Face& face). By 'adding' a new column for the EndpointID, do I set a new method NextHop(Face& face, endpoint) or "update" it to have the endpoint column fixed?

Also, I see that there are multiple endpoint type, i.e. protocol::endpoint, udp::Endpoint, tcp::Endpoint. How should I handle these and what should be the default type I use for the EndpointID column in the NextHop table?

Currently the addNextHop(Face& face, uint64_t cost) method says: If a NextHop record for face already exists its cost is updated.

However, should we not also check if the a certain endpointId also exists before we create a new entry and/or update the cost? i.e. I think it should be updated to addNextHop(Face& face, uint64_t endpointId, uint64_t cost), and perform checking if an endpoint id to a face exists or not.

Basically every occurrence of “Face” is replaced with a tuple of Face+EndpointId.
A notable exception is how FIB and PIT reacts to a face being destroyed: in that case, relevant entries with the affected face and any endpointId are removed.

Junxiao Shi wrote:

Basically every occurrence of “Face” is replaced with a tuple of Face+EndpointId.

Meaning in the Pit::erase, I need to set "Face -> "Face + endpointId" ?

A notable exception is how FIB and PIT reacts to a face being destroyed: in that case, relevant entries with the affected face and any endpointId are removed.

Does this mean when a simple remove- Fib::removeNextHop(Entry& entry, const Face& face) is called (meaning the default endpointId=0), this should remove all entries regardless of the endpointId value?