NLSR

Design question:

Why is LSDB Status dataset designed to have only one element: the LsdbStatus block?

Why not put AdjacencyLsa, CoordinateLsa, NameLsa directly in the dataset?

Junxiao Shi wrote:

Design question:

Why is LSDB Status dataset designed to have only one element: the LsdbStatus block?

Why not put AdjacencyLsa, CoordinateLsa, NameLsa directly in the dataset?

Maybe I am misunderstanding the question, but LsdbStatus, a list of AdjacencyLsas, a list of CoordinateLsas, and a list of NameLsas are all separately retrievable.

/$ROUTER_NAME/lsdb/list
/$ROUTER_NAME/lsdb/adjacencies
/$ROUTER_NAME/lsdb/coordinates
/$ROUTER_NAME/lsdb/names

Did you mean why have the LsdbStatus block at all and instead just directly publish the lists in the dataset?

To implement Junxiao suggestion, you can do something similar to (not the exact code):

Face face;
face.registerPrefix("/$ROUTER_NAME/lsdb/",
                [&face] (const Name& prefix) {
                  face.setInterestFilter(Name(prefix).add("adjacencies"),
                                         onAdjacenciesCallback);
                  face.setInterestFilter(Name(prefix).add("coordinates"),
                                         onCoordinatesCallback);
                  ...
                },
                [] (const Name& prefix, const std::string& errMsg) {
                  ... log errors
                });