20160107 conference call talked about long term plan for future releases. We are considering the following categories:
performance¶
Our target scenario is big science data.
We can look into issues including: congestion control, table lookup, ContentStore, repo-ng.
We should spend serious effort to benchmark current performance, find out where's the slowdown, understand current performance, have a plan on how to improve.
mobile and embedded¶
Currently NFD works on home routers and mobile phones, but we haven't put enough effort on those platforms.
Note: home routers and mobile phones are capable in terms of computing resources; they are not constrained devices.
forwarding behavior and strategy¶
We should look at strategy from scratch, and make it more powerful and more flexible.
We should investigate the relation between recently proposed "route announcement annotation" and best-route strategy.
Also, we should investigate the interaction between LinkService's loss detection and limited recovery (depends on sequence number), and strategy's loss detection (needs timeout).