What is Secure Multi-Party Computation (MPC)?

A protocol that lets several parties jointly compute a function over their combined data while each keeps its own input private — "analyze a shared database you can't see."

* MPC: parties compute on secret shares; only the result is revealed, raw inputs never shared. *

* Yao's Millionaires: learn who is richer without revealing either wealth. *

No party reveals its raw data; instead, inputs are cryptographically split (e.g. into secret shares) and the parties exchange messages so that, at the end, everyone learns only the agreed output and nothing about each other's inputs.

The canonical illustration — Yao's Millionaires' Problem: two millionaires want to know who is richer without revealing how much each has. MPC outputs just the comparison result.

Key properties:

• Inputs stay decentralized (no trusted central party needed).
• Only the final result is revealed.
• It's interactive — parties exchange many rounds of messages.

Go deeper:

• Secure multi-party computation (Wikipedia) — adversary models and protocol families (secret sharing, garbled circuits).
• Yao's Millionaires' problem (Wikipedia) — the canonical illustration with the actual comparison protocol.