The Transaction Graph
A weighted, directed graph capturing economic relationships, used via snapshot-relative diffusion to allocate incentives and resist Sybil attacks.
The transaction graph is a weighted, directed graph that captures economic relationships between participants. Each completed interaction adds or updates an edge, and the protocol uses the resulting connectivity (via snapshot-relative diffusion) to allocate incentives and resist Sybil manipulation.
Graph Structure
Let the global transaction graph at epoch
: participants (buyers, producers, agents, domains, etc.) : directed edges representing completed interactions : nonnegative edge weights
In a marketplace setting:
- buyer
purchasing from producer adds edge - an optional reverse edge
can represent fulfillment confirmation, dispute outcomes, or service-proof acknowledgements
Edge weights
Each completed transaction produces an edge weight:
Where:
amountis the economic value (price, fee base, etc.)qualityaccounts for dispute outcomes, refunds, chargebacks, delivery SLAs, etc.proof_factoris derived from attached service proofs and identity proofs
The protocol constrains weights to prevent pathological abuse (per-edge min/max, per-transaction caps, epoch caps, and/or decay).
Key Features
1. Dynamic Adjustments
The transaction graph dynamically adjusts based on participant interactions. As transactions occur, edge weights are updated, causing changes in connectivity and node influence. This creates a self-optimizing system where token distributions reflect the evolving state of the network.
2. Connectivity as a Measure of Value
The graph not only captures transaction volume but also connectivity:
- Nodes with more connections to well-connected nodes are considered more influential.
- This approach ensures that participants contributing to network growth through broad connectivity earn higher rewards, not just participants with high transaction volumes with a single counterparty.
3. Sybil Resistance
The graph’s structure inherently resists manipulation through Sybil attacks:
- Sybil nodes (fake users) typically form isolated clusters without strong connections to real nodes.
- The graph's weighting system prioritizes connections that enhance network-wide connectivity, making it difficult for isolated Sybil nodes to earn high rewards.
Next Steps
The transaction graph sets the foundation for diffusion and verification: