Technical definition
Multi-party computation (MPC) is a subfield of cryptography that enables a set of participants to collaboratively compute a function over their private inputs while ensuring that no party learns anything about the other parties' inputs beyond what is revealed by the function's output. This is achieved through protocols like secret sharing, garbled circuits, or homomorphic encryption, which allow computations to be performed on encrypted or split data. MPC ensures privacy, correctness, and fairness in distributed systems where trust is limited or absent.
An analogy
Imagine three people who each hold a piece of a puzzle. Instead of showing their pieces to each other, they work together to assemble the full puzzle by describing their pieces' shapes and positions without revealing the actual image on the pieces. The final puzzle is complete, but no one learns what the other pieces looked like.
Why it matters
MPC is critical in scenarios where sensitive data must be processed without exposing it to any single party, such as secure auctions, privacy-preserving data analysis, or decentralized identity verification. It enables collaborative computation without trust in a central authority.
Where it came from
The concept of MPC was introduced in the 1980s by Andrew Yao, who proposed the 'millionaire problem'—a scenario where two parties can determine who is richer without revealing their actual wealth. Early protocols like Yao's garbled circuits laid the groundwork for modern MPC systems.
Example in practice
In blockchain, MPC is used for threshold signatures, where multiple parties (e.g., wallet holders or validators) jointly sign a transaction without any single party knowing the full private key. This is employed in protocols like DFINITY's Internet Computer or some enterprise blockchain solutions.
Common misconceptions
- MPC eliminates the need for trust entirely → Correction: MPC reduces trust requirements but may still rely on assumptions like honest-but-curious adversaries or secure communication channels.
- MPC is only useful for blockchain → Correction: MPC has applications in finance, healthcare, and machine learning, such as secure multi-party auctions or privacy-preserving genomic research.
Risks & limitations
- Computational overhead: MPC protocols often require significant computational resources and bandwidth, making them impractical for large-scale or real-time applications.
- Side-channel attacks: Poorly implemented MPC systems may leak information through timing, power consumption, or other side channels.
- Trusted setup requirements: Some MPC protocols (e.g., those using garbled circuits) may require a trusted setup phase, which can introduce vulnerabilities if compromised.
- Denial-of-service risks: Malicious parties may disrupt MPC protocols by refusing to participate or sending incorrect inputs, leading to failed computations.