Codex System Analysis #177
Description
- is the slot model and assumption sound? we need someone to sit down and get numbers on this #178
- maybe we do need to switch it and go the other direction w/ node-level assumptions #179
- can we treat all the slots in a node as one? if one test failed then all slots are failing, does this align w/ node-level failure rather than slot-level failures? #180
- is there a better way to do this than having these two extremes? #181
- can we keep slot-level assumption and do aggregation and is that sound? not even in the erasure-coding dimensions. is it a sound assumption that we're storing things in slots and many slots can be in one node. #182
- how frequent should the proof requests be, how many times does a node need to send proofs a day and based on what assumptions? #183