The world’s first commercial implementation of Multi-Party Computation (MPC) took place in 2008: the Danish Sugar Beet Auction. Here is the story of how sugar beets in Denmark helped pioneer this technology.
MPC is a cryptographic technique that first emerged in the 1980s. For many years it remained theoretical, a matter of hypothetical scenarios. That changed in 2008, when the first commercial implementation ran as a double auction for Danish sugar beet contracts.
When the European Union drastically reduced its support for sugar beet production, a double auction was needed to set the price of contracts in Denmark. MPC ran the auction, keeping the farmers’ bids confidential while still producing a fair outcome. One of the people behind the breakthrough was Ivan Damgård, Partisia’s co-founder.
The demonstration opened the door for real-world applications of MPC, proving it could move beyond theory and solve business problems involving data that has to stay confidential.
Background: why sugar beets?
Denmark has a significant sugar beet industry, with around 4,000 farmers growing beets for Danisco, a Danish producer of food ingredients and other consumer products.
Each farmer holds a contract specifying how many beets they may produce and the price Danisco will pay. Those contracts can be traded between farmers, but the process had traditionally relied on limited one-to-one negotiation.
The turning point came when the EU significantly reduced its financial support for sugar beet production. Farmers came under immediate pressure to reallocate contracts to whoever could produce most profitably, which made finding the market clearing price, the point where demand meets supply most effectively, crucial. A double auction was the right instrument for the job.
A double auction needs an auctioneer, and nobody could be one
In a double auction, buyers and sellers simultaneously submit bids to buy and asks to sell. Pricing of this kind is most effectively managed through an auctioneer, who keeps competition fair and the process efficient.
That raised the obvious question: who should be the auctioneer?
- Danisco? No. Bids reveal private information about a farmer’s economy, which Danisco could misuse.
- The farmers themselves? Also no. Farmers do not fully trust each other, and Danisco wanted some control, since contracts can serve as security for debt owed to the company.
- An external consulting firm? Then everything depends on that single party keeping the data confidential and its tools working flawlessly.
“The problem with a double auction is that it requires a trustworthy and coordinating third-party. One that would protect each farmer’s confidential information while still determining fair prices and optimal outcomes for all.”
Kurt Nielsen, CEO and Partner at Partisia
What do you do when no single party can be trusted with sensitive data? MPC removes the need for a middleman, and the stage was set for its first commercial use.
The solution: a virtual auctioneer
The role of the auctioneer was played by MPC, run between three parties: Danisco, DKS (the Danish sugar beet growers association), and the SIMAP project (Secure Information Management and Processing, sponsored by the Danish National Research Agency).
A three-party solution was chosen partly because it was natural given the three main stakeholder groups, and partly because it allowed the use of efficient information-theoretic tools such as secret sharing.
“When no single party can be trusted with sensitive data, Multi-Party Computation steps in as an ideal solution. It allows multiple stakeholders, who may have competing interests, to compute a result together without exposing their individual inputs.”
Kurt Nielsen, CEO and Partner at Partisia
Why MPC suited the job
No party sees all the data. With MPC, nobody sees the bids in unencrypted form, which removes the need for any single party to carry responsibility for keeping them private.
Privacy is ensured. Because the data is split and processed so that no one holds all of it at once, the individual bid amounts stay protected.
Collaboration gets easier among conflicting interests. Different stakeholders often have clashing goals. MPC gives them a built-in way to work together securely without negotiating a complicated security arrangement.
One secure hardware device is not enough. Relying on a single piece of hardware creates a single point of trust, and of failure, that everyone must depend on. MPC does not rest on one device.
The result
The auction determined the market clearing price with three servers computing on encrypted bids that were never decrypted, so no one could see any farmer’s individual bid.
A survey among the farmers afterwards showed what it meant to them:
- 81% found that MPC simplified the process of trading contracts.
- 78% said confidentiality of bids was important.
- 86% were satisfied with the level of confidentiality provided.
All parties were extremely happy with the solution, so the auction ran every year until 2015, when it was no longer needed.
What the Danish Sugar Beet Auction teaches
It proved that MPC works in real commercial settings. It showed that private information from many participants can be combined without ever revealing anyone’s individual data, which creates room for collaboration even where participants have competing interests and sensitive information to protect. Those are exactly the challenges that often block innovation and growth.
Curious what confidential computation could unlock in your own industry?
