Overview of the APDB protocol Sample Clauses

Overview of the APDB protocol. For the PD subprotocol with identifier ID, it has a simple structure of four one-to-all or all- Store Stored Lock Locked to-one rounds: sender −→ parties −→ sender −→ parties −→ sender. Through a Store message, every party Pi receives store := ⟨vc, mi, i, πi⟩, where mi is an encoded fragment of the sender’s input, vc is a (deterministic) commitment of the vector of all fragments, and πi attests mi’s inclusion in vc at the i-th position; then, through Stored messages, the parties would give the sender “partial” signatures for the string ⟨Stored, ID, vc⟩; next, the sender combines 2ƒ + 1 valid “partial” signatures, and sends every party the combined “full” signature σ1 for the string ⟨Stored, ID, vc⟩ via Locked messages, so each party can deliver lock := ⟨vc, σ1⟩; finally, each party sends a “partial” signature for the string ⟨Locked, ID, vc⟩, such that the sender can again combine the “partial” signatures to produce a valid “full” signature σ2 for the string ⟨Locked, ID, vc⟩, which allows the sender to deliver done := ⟨vc, σ2⟩. For the RC subprotocol, it has only one-round structure, as each party only has to take some output of PD subprotocol as input (i.e., lock and store), and multicasts these inputs to all parties. As long as an honest party inputs a valid lock, there are at least ƒ + 1 honest parties deliver valid stores that are bound to the vector commitment vc included in lock, so all parties can eventually reconstruct the dispersed value that was committed in the commitment vc. Algorithm 1 Validation func of APDB protocol, with identifier ID function ValidateStore(i′, store): D ValidateStore verifies store commits a fragment mi in vc at i-th position