Breaking Silos: The First Proxy-Free Federated Crowdsourcing via Blockchain

TL;DR

Current crowdsourcing platforms like Amazon Mechanical Turk operate as isolated islands, where sensitive task and worker data are exposed to centralized brokers. This paper proposes a breakthrough: a Federated Crowdsourcing System that allows different brokers to share resources (tasks and workers) securely. By utilizing Blockchain, Rewritable Deterministic Hashing (RDH), and Puncturable Encryption, the authors eliminate the need for a "Trusted Third Party," making the system truly decentralized and "proxy-free."

The Core Motivation: The Trust Paradox

In the current landscape, if Broker A wants to let Broker B's workers solve its tasks, it would traditionally need a central authority to manage encryption keys. However, in a competitive market, finding a third party that everyone trusts—and who won't leak sensitive task specifications—is nearly impossible. Furthermore, once an authorization is revoked, how do you prevent a former partner from reading the "permanent" records on a transparent blockchain?

Methodology: The Cryptographic Trio

The authors solve the trust paradox using three pillars:

1. Federated Identity via RDH

The system uses Rewritable Deterministic Hashing (RDH) to enable cross-broker search. Instead of sharing a master key, Broker $b_i$ generates an authorization ticket $A_{b_i\to b_j}=g^{F_{b_i}^1/F_{b_j}^2}$ for Broker $b_j$. This allows the Smart Contract (SC) to transform a search query from $b_j$ into a valid trapdoor for $b_i$'s encrypted index without the SC or $b_j$ ever seeing the underlying keywords.

2. Guarding the Ledger with Puncturable Encryption

Blockchain data is immutable and public. To handle revocation, the authors use Puncturable Encryption. When Broker $b_j$'s access is revoked, Broker $b_i$ "punctures" the key space by adding $b_j$'s unique tag to a revoked list. New task indices are encrypted specifically to exclude these punctured tags, ensuring that even if $b_j$ still has its old keys, it cannot decrypt any new data.

3. Model Architecture

The architecture moves the heavy lifting of task-worker matching to the Smart Contract, ensuring the process is verifiable and correct.

Experimental Results: Is it Practical?

The authors implemented the prototype on Ethereum. A critical concern with blockchain is usually the Gas Cost and Latency.

• Latency: The local processing time for 300 indices is under 5 seconds, meaning the bottleneck remains the block mining time (~1s in the experiment), not the cryptography.
• Scalability: The system scales linearly with the number of brokers. Even with 14 brokers, authorization confirmation on-chain stays under 4 seconds.
• Cost: At a gas price of 1 Gwei, authorization costs roughly $0.07USD,andpostingtaskindicesisabout$0.19 USD. This is remarkably affordable for enterprise-level crowdsourcing.

Critical Analysis & Conclusion

The true value of this work lies in achieving Proxy-Free operation. Most "decentralized" papers still hide a "trusted admin" in the shadows for key management. This paper kicks the admin out of the room.

Limitations: Currently, the system primarily supports exact keyword matching. For real-world crowdsourcing, stakeholders often need Range Queries (e.g., "Find workers within 5km") or Fuzzy Matching. The authors acknowledge this as the next frontier for their research.

Takeaway: By bridging advanced cryptography (Puncturable Encryption) with distributed ledgers, we can transition from "isolated silos" to a "unified global market" of collective intelligence, all while maintaining strict user privacy.