Abstract
Smart contracts are inherently isolated from real-world data, necessitating the use of oracles. However, traditional centralized oracles introduce single points of failure. Infochain addresses this by utilizing game-theoretic peer-consistency mechanisms to elicit truthful data from self-interested contributors on the Ethereum blockchain. By implementing novel smart contract optimizations, Infochain overcomes the high gas costs typically associated with complex on-chain incentive structures, making decentralized data elicitation economically viable even when external incentives to lie exist.
Introduction
The utility of smart contracts is often limited by their inability to verify external events. While oracles bridge this gap, centralized solutions contradict the decentralized nature of blockchain technology. Crowdsourcing data is an alternative, but it faces the "outside incentive" problem: if a contract triggers a payout based on a specific report, users are incentivized to report that outcome regardless of the truth. Furthermore, the computational overhead of verifying these reports on-chain can be prohibitively expensive. Infochain proposes a decentralized, trustless framework that aligns user incentives with truth-telling while minimizing operational costs.
System Architecture and Workflow
Infochain operates through a structured interaction between information requesters and providers, mediated by a smart contract. The process is designed to prevent collusion and "copycat" behavior through a cryptographic commit-reveal protocol.
The workflow consists of five primary stages:
- Deployment: The requester deploys the smart contract with a budget and deposit, posting the questions to be answered.
- Selection: Information providers select questions to answer.
- Commitment: Providers submit cryptographic hashes (commitments) of their answers along with a deposit to ensure participation.
- Revelation: After the commitment phase, providers reveal their actual answers and secret keys.
- Settlement: The contract processes payments and returns deposits based on the chosen peer-consistency mechanism.
Peer-Consistency Mechanisms
To determine the truth without a ground-truth source, Infochain employs peer-consistency mechanisms. These mechanisms reward users based on the statistical correlation of their reports with those of their peers. The system evaluates three primary models:
- Output Agreement (OA): Rewards simple agreement between peers.
- Dasgupta and Ghosh (DG): A more complex mechanism that handles multi-task settings.
- Peer Truth Serum for Crowdsourcing (PTSC): Rewards "surprising" agreement, making honesty the most profitable strategy even in the face of external bribes.
On-Chain Optimizations
A critical contribution of Infochain is its focus on gas efficiency. On Ethereum, storage operations (writing to the blockchain) are significantly more expensive than computation.
Bit-Packing
To minimize storage costs, Infochain uses bit-packing to store multiple binary answers within a single 256-bit Ethereum word. By treating the word as a bit vector, the system can store up to 42 answers simultaneously. This approach results in a near-constant writing cost regardless of the number of tasks, as shown in the optimization analysis.
Memory Management
The system further reduces costs by precomputing shared values—such as report frequencies—in memory during execution. By avoiding intermediate writes to the blockchain's permanent storage, Infochain significantly lowers the total gas consumption for complex calculations.
Evaluation and Results
The researchers evaluated Infochain using a real-world Quality of Service (QoS) dataset containing response times for thousands of web services. Simulations involving honest, random, and adversarial agents demonstrated that the PTSC mechanism effectively incentivizes truth-telling.
Key findings include:
- Economic Viability: For a typical service with 95% uptime, the cost to ensure truthfulness is approximately 1.2% of the potential refund amount.
- Scalability: The optimized implementation maintains low writing costs as the number of tasks increases, whereas baseline costs scale linearly.
- Mechanism Efficiency: OA and PTSC were found to be more gas-efficient than the DG mechanism, which requires more complex tracking of peer interactions.
Conclusion and Future Work
Infochain demonstrates that decentralized oracles can be both secure and economically feasible. By combining game theory with low-level smart contract optimizations, it provides a robust framework for trustless data elicitation. Future research will explore extending these mechanisms to continuous data spaces and utilizing Layer 2 scaling solutions to further enhance performance and reduce costs for larger peer groups.
Summary updated