codex-research issues

Zero knowledge proofs

epic

Milestone 17 - Research paper about Codex objectives, design, and analytical model projections

2

- identify conferences to publish academic publication results of Codex Proposed structure: * Abstract * Introduction * State of the art - IPFS - Filecoin - Storj - Sia -...

leobago

epic

how frequent should the proof requests be, how many times does a node need to send proofs a day and based on what assumptions?

jessiebroke

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.

jessiebroke

is there a better way to do this than having these two extremes?

jessiebroke

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?

jessiebroke

maybe we do need to switch it and go the other direction w/ node-level assumptions

jessiebroke

is the slot model and assumption sound? we need someone to sit down and get numbers on this

jessiebroke

capture top-level folder structure into README (and use standard naming)

1

Added the info @dryajov provided me on Discord about top-level structure. Suppose it might be useful for newcomers or casual perusers.

gmega

Multi-threading

1

Jaremy will have write-up under considerations and approaches and actual user stories will shake out of that. Good to start w/ IO file system access async. Investigate the threadpool approach,...

jessiebroke

epic

codex-research
codex-research copied to clipboard

Metadata

Zero knowledge proofs

Milestone 17 - Research paper about Codex objectives, design, and analytical model projections

how frequent should the proof requests be, how many times does a node need to send proofs a day and based on what assumptions?

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.

is there a better way to do this than having these two extremes?

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?

maybe we do need to switch it and go the other direction w/ node-level assumptions

is the slot model and assumption sound? we need someone to sit down and get numbers on this

capture top-level folder structure into README (and use standard naming)

Multi-threading

← Metadata

Owner

Metadata

codex-research codex-research copied to clipboard

Metadata

← Metadata

Owner

Metadata

codex-research
codex-research copied to clipboard