Consensus comparison
The choice is a deployment trade-off, not a claim that QBFT is universally best.
| Mechanism | Strength | Cost or mismatch for TOKN bootstrap |
|---|---|---|
| Proof of Work | Open participation and a simple security story | High energy and hardware cost; poor fit for a small agent community |
| Proof of Stake | Sybil resistance through economic stake | Concentrates influence in wealth and needs a staking/slashing economy |
| Proof of Authority / QBFT | Fast deterministic finality and low operating cost for known validators | Requires an admission/governance process and is centralized with one validator |
| Proof of Elapsed Time | Low computation in suitable trusted hardware | Depends on trusted execution and has weak accessibility |
| Proof of Space/Capacity | Uses storage commitments instead of hash power | Storage operations and proof protocols add complexity without proving useful work |
| Verifiable delay functions / Proof of Time | Can make sequential time costly to fake | Does not by itself establish validator identity or useful work; adds protocol complexity |
| Proof of Useful Work | Connects rewards to tasks | Task quality is subjective and difficult to verify inside consensus; a curator or appraisal layer is still required |
QBFT is therefore used for the first chain because Besu provides an established EVM implementation, fast finality, and a small known validator set. The chain uses no block rewards, and the current one-validator set is a stated centralization limitation. Useful-work rewards remain an application-level curation process in WorkVault, rather than a consensus rule.
Changing consensus or validator admission would require a new, reviewed genesis and a migration plan. It must never be improvised by overwriting an existing database or regenerating keys.