Metis Computing Services
Metis is a platform that supports privacy computing and aims to process data efficiently and securely while protecting privacy. Privacy computing is a technology that processes and analyzes data in a form that encrypts or otherwise protects its privacy. Metis supports two different types of privacy computing methods: Secure Multi Party Computation (SMPC) and Privacy Outsourcing Computation.
Secure Multi party Computing (SMPC)
Secure multi-party computation is a research field in cryptography that aims to allow multiple parties to jointly complete data computation tasks without leaking their respective inputs. In SMPC, each participant holds a private copy of the data, but they need to collaborate to calculate certain aggregation characteristics of the data, such as summation or averaging. The entire calculation process is carried out in an encrypted state, ensuring that each participant cannot access the data of others. The secure multi-party computation protocol is used for privacy computation between computing nodes, and the computation results are returned to the computing result party through blockchain smart contracts. If it is an AI model training, the trained AI model can be deployed to Layer 3’s AI network as an AI agent and published as a service through the RELOAD network.
In the Metis platform, SMPC can be implemented through the following methods:
-
Cipher calculation: All data provided by the participating parties exist in ciphertext form. During the calculation process, all parties cooperate, but only decrypt the parts that need to be calculated. After the calculation is completed, the results are encrypted and shared with all parties again.
-
Protocol and Algorithm: Metis uses standardized protocols and algorithms to ensure the security of the computing process.
-
Threshold scheme: In order to improve the fault tolerance of the system, Metis may adopt a threshold scheme, which means that as long as no more than half of the participating parties commit wrongdoing, the entire calculation process can still be safely completed.
Privacy Outsourcing Computation
Privacy outsourcing computing refers to users outsourcing computing tasks to third-party service providers (such as cloud servers), while requiring the third party to complete the computation without leaking data. In this model, user data is encrypted before being sent to a third party, who is only responsible for computation and is not aware of the specific content of the data. If users have their own data and algorithms but do not have sufficient computing power, they can outsource their data (homomorphic encrypted) and algorithms to third-party computing nodes for computation. Data and algorithms can be distributed to multiple computing nodes for parallel computing, and computing tasks can be decomposed based on data or models. After the calculation node completes the calculation, it returns the calculation result and calculation proof to verify the correctness of the calculation.
The Metis platform supports privacy outsourcing computing in the following ways:
-
Verifiable calculation: Users encrypt the functions and input data that need to be calculated and send them to the service provider. After completing the calculation, the service provider returns the calculation result and verifiable proof of the result. Users can verify the proof to ensure the correctness of the calculation results without needing to understand the detailed calculation process.
-
Encryption algorithm: Metis uses advanced encryption algorithms to protect user data and maintain its privacy during data processing and analysis.
-
Zero knowledge proof: In some cases, Metis may use zero knowledge proof techniques to enable service providers to verify the correctness of calculations without needing to know any additional information about the data or calculations.