Skip to content

weighted-voting.md

Latest commit

 

History

History
112 lines (84 loc) · 5.33 KB

weighted-voting.md

File metadata and controls

112 lines (84 loc) · 5.33 KB
Raw

Weighted quorum voting

Quorum slices assign each attestor an integer trust weight from 1 through 100. A weight is relative within one slice; it is not a global reputation score. Slice creators should assign weights from documented evidence and avoid giving one organization unilateral control unless that is an explicit policy decision.

Consensus algorithm

For a credential and slice, the contract evaluates each unexpired, positive attestation that was submitted through that slice. Suspended attestors contribute zero. Each accepted attestation contributes the attestor's weight captured at submission time:

achieved_weight = sum(captured_weight(attestation))
consensus       = achieved_weight >= required_weight

Capturing the weight at submission prevents a later weight update from revaluing past attestations. New attestations use the current weight. Threshold changes intentionally apply to future consensus queries so credential types can adopt different policies over time.

The slice creator chooses one of two threshold modes:

  • Absolute: required_weight = threshold, where 1 <= threshold <= total_weight.
  • Percentage: required_weight = ceil(total_weight * percentage / 100), where the percentage is from 1 through 100. Ceiling division prevents rounding from weakening the configured policy.

For weights [2, 1, 1], an absolute threshold of 3 is met by the weight-2 licensing body and either weight-1 attestor. A 67% threshold requires ceil(4 * 67 / 100) = 3 weight units.

Dynamic management and security

Only the slice creator can add or remove attestors, change weights, or change thresholds. Weight updates outside 1-100 are rejected. An absolute-threshold slice also rejects a weight reduction that would make its threshold unreachable. Percentage thresholds automatically track the current total weight.

get_weight_audit provides the immutable sequence of weight changes. get_weight_distribution returns attestor count, total, minimum, maximum, and integer-average weight. Consensus history records achieved, required, and total weight when a decision reaches threshold.

Assignment guidance

  • Use a small documented scale, such as 1 for direct employers, 2 for accredited institutions, and 3 for national regulators. The 1-100 range is a limit, not a reason to use all values.
  • Prefer a threshold that requires independent trust domains. A simple majority is above 50%; a common supermajority policy is 67%.
  • Review concentration through get_weight_distribution. If maximum_weight >= required_weight, one attestor can decide consensus alone.
  • Change weights prospectively when evidence changes. Revoke or expire an attestation when its underlying evidence, rather than the attestor's general trust level, is no longer valid.

Economic Security Analysis

QuorumProof provides tools to quantify and mitigate the cost of attacking a slice's consensus threshold. See docs/economic-security-model.md for the complete formal model.

Attack Cost Estimation

Query get_slice_attack_cost_estimate to understand your slice's vulnerability:

attack_cost = get_slice_attack_cost_estimate(slice_id)

Returns:

  • corrupt_existing_cost: Minimum cost to corrupt existing attestors for threshold
  • sybil_attack_cost: Cost to launch a Sybil attack
  • concentration_risk_score: 0-100, where 100 is maximum vulnerability
  • has_single_point_of_failure: true if maximum_weight >= required_weight
  • attack_detection_probability: Likelihood attacker is detected

Reputation-Tied Weighting Mitigation

If a slice detects the concentration risk score is too high, enable reputation-tied weighting:

set_reputation_weighting_enabled(creator, slice_id, enabled: true)

When enabled:

  • Effective weight = base_weight × (reputation_score / 100)
  • Corrupted attestors face reputation penalties, reducing their effective weight
  • Sybil attackers cannot immediately vote at full weight
  • Attack costs increase 2-10x for representative configurations

See docs/economic-security-model.md for attack cost multipliers and configuration guidance.

Monitoring Attestor Security

Query get_slice_security_profiles to monitor each attestor's trustworthiness:

profiles = get_slice_security_profiles(slice_id)

Returns for each attestor:

  • weight: Base voting weight
  • reputation_score: Current reputation (0-100, 100 is fully trusted)
  • confirmed_malicious_count: Number of proven equivocations

API summary

  • create_slice: create a slice with an absolute weight threshold.
  • create_slice_percentage: create a slice with a percentage threshold.
  • update_attestor_weight: update one attestor's prospective weight (creator only).
  • update_slice_threshold: switch to or update an absolute threshold (creator only).
  • update_percentage_threshold: switch to or update a percentage threshold (creator only).
  • get_slice_threshold_config: inspect mode, configured value, and effective required weight.
  • get_weight_distribution / get_weight_audit: inspect metrics and change history.
  • get_slice_attack_cost_estimate: estimate economic cost to attack slice consensus.
  • get_slice_security_profiles: monitor reputation and trustworthiness of attestors.
  • set_reputation_weighting_enabled: enable/disable reputation-tied weighting (creator only).
  • get_effective_weight: query effective weight of attestor considering reputation.