Secret Sharing and Multi-Party Computation: The key to secure data collaboration

It’s one of the methods we at Partisia use to keep our partners’ data safe, whether it’s private information or confidential business secrets.

If we also want to compute this data, we use Multi-Party Computation (MPC) technology to perform calculations on the distributed data, without ever exposing the underlying sensitive information. 

To understand how Secret Sharing and MPC work together, first let’s break down each technique on its own.

Secret Sharing is a cryptographic technique for taking a piece of private information or secret and splitting it up into multiple smaller parts, known as shares. 

These shares are then distributed among different parties. On their own, individual shares reveal nothing, but when enough shares are combined, they reconstruct the original secret.

This ensures that no single party can reveal or misuse the entire secret, which enhances privacy and security even when participants may not fully trust one another.

Multi-Party Computation (MPC) is a cryptographic technique that allows multiple parties to compute a function together and securely over their private inputs without ever revealing the inputs to each other.

No participant learns another’s private data; only the final results of these computations are revealed.

Secret Sharing is what drives Multi-Party Computation. It is the ability to split a piece of data into several coded parts (“secret shares”) and distribute them across servers in such a way that no single server learns the secret but together, the servers can perform computations on this securely shared information.

This synergy protects sensitive information while still enabling collaborative data processing.

To understand even more how they work together, let's look at a business example where three companies calculate their average revenue based on their quarterly revenues while keeping their individual figures secret.

Companies A, B, and C each have a list of private numbers – say, their quarterly revenues.

Goal: The companies want to calculate the average of those revenues, without revealing their individual figures to each other.

Each company splits its private numbers into random-looking “shares”.

Those shares are then distributed to the other companies in such a way that no single company (by looking at just the shares it received) can guess what the original revenue numbers of the others are.

Once each company has received the “shares” of everyone else’s data, they use Multi-Party Computation (MPC) to perform the calculation.

Instead of exposing their private numbers, each company works exclusively with the random-looking shares it holds. 

Nobody learns each other’s input data; they only learn the final combined result they all agreed to calculate – their average revenue. 

With Secret Sharing and Multi-Party Computation, private data stays protected even if malicious actors try to compromise one of the parties or if hackers breach a system.

Let’s explore two common scenarios: compromising attempts from within and external cyber attacks and see how this technology keeps sensitive information safe.

If company A tries to peek at the shares it received from B and C, it will only see random-looking data. Even if A tries to combine those random shares with its own data, they still won’t reveal any information about B’s or C’s private numbers. That’s because each share is indistinguishable from random data unless enough shares are brought together under the rules of the protocol.

Even if hackers break into one of the companies’ systems, say company A’s system, they only find A’s copy of the secret shares from the other companies, not the original data. 

These shares, on their own, are worthless for reconstructing the original information. Without the matching shares from B and C, the hacker can’t uncover any private data, keeping everyone’s information safe from unauthorized exposure.

Secret Sharing and Multi-Party Computation (MPC) work hand in hand to protect private data while performing calculations to get insights from previously siloed data.

By splitting a secret into shares, no one party can see the full picture on its own. When we go one step further with MPC, parties can compute results without revealing the underlying data to anyone.

• Secret sharing is like handing out pieces of confidential data in a scrambled way; MPC is the process by which businesses can perform calculations and collaborate using just the scrambled pieces. 

• This technology combination is useful because no one sees anyone else’s data and you can still work together to uncover insights from “secret” or siloed data.

• The method ensures that no single party can piece together the full picture of private information, protecting the raw data from privacy breaches or insider threats.

Are you ready to enter the next level of secure data sharing?

Our whitepaper with an introduction to Confidential Computing on the Partisia Platform is exactly what you need to get started.