- Subscribe OK needs the HEAD
- Define the client-facing fetch API (track, group ID, object ID range)
- Implement the fetch retrival and output at line rate.
- Client fetches.
- Fetches to relays that have no data in a cache, should go upstream
- Relays should cache incoming fetch data in the same way that it would cache any incoming data.
- LRU cache.
- We should use the memory cache for both the cache and the memory for the queue.
- How are the cache output priorities / transport work.