NFD

20140512 research meeting with Tian Song decides that Tian will provide advice on how to implement the trie for RIB.
One possibility is to implement RIB on top of NameTree, if class NameTree can provide sufficient algorithms for RIB usage. NameTree entry will need a generic type to support either FIB/PIT/Measurements/StrategyChoice or RIB.

Anyone can share a copy of the design document for NameTree? I don't know enough about it to make a comment on its applicability in RIB at this moment.

Please find the data structure design of the NameTree on Slide 11 of this document (https://dl.dropboxusercontent.com/u/2661516/14_02_14_Table_Design_updated.pdf), which was used when we discussed the NameTree design. I have updated the slides with the implemented APIs.

A more general description of NameTree and its functions (prepared by Junxiao and Alex?) is available at http://redmine.named-data.net/issues/1187.

Lan Wang wrote:

Anyone can share a copy of the design document for NameTree? I don't know enough about it to make a comment on its applicability in RIB at this moment.

Is the current NameTree implementation generic enough for us to store RIB in it? Alex told me that it is not right now.

NameTree entry needs to be extended / generalized to support RIB.
NameTree algorithms should be sufficient.

There is a long term benefit to have RIB and NFD tables both use NameTree, rather than creating a separate structure, because any improvement to the NameTree can benefit both.

I wonder if Vince is able to modify the implementation of NameTree to support generic entries and make it implement an interface (which can be implemented by other data structures, e.g., trie).