General Architectural Outline

The design of Corex Network is highly adaptable and dynamic, catering to a wide range of distributed confidential computing requirements. Each confidential computing task is divided into distinct computation units, which are carried out within the joint computing environments.

Joint computing environments function within groups of nodes called clusters. In traditional computing terms, clusters serve a similar role to system hardware, while the joint computing environments represent the system’s state. Computation definitions are comparable to functions, with each scheduled computation representing an instance of a function call.

Below you can find the essential concepts and definitions:

Nodes:

Within our network, a node is a distinct computing service provider that participates in distributed computation tasks. Nodes can engage in multiple computations simultaneously and for each task, the node manages a portion of the data being processed, specifically holding a keyshare of the encrypted data.

These nodes work together to compute on their data fragments while ensuring that no individual node can view the full data held by others.

Clusters:

Clusters are collections of nodes assembled by users to perform computational tasks. A cluster acts as the group of nodes that collaboratively carries out a multi-party computation.

Computations:

A computation refers to a specific task to be processed by an joint computing environment . It is initiated by a computation client and includes a description of the circuit to be executed, the entities authorized to oversee the execution, methods for handling input and output data, and callback functions with additional attributes that can be included as well.

Computations can be linked together in sequences, where the result of one task is passed to the next through callbacks, forming a chain of dependent operations.

Multi Party Computation:

Multi-Party Computation is a cryptographic method that permits several participants to collaboratively compute a function using their own inputs, while ensuring those inputs remain hidden from each other. MPC allows the parties to generate a collective outcome without disclosing their personal data, maintaining both privacy and security throughout the computational process.

Encrypted Channels:

An encrypted channel refers to a communication route that ensures the privacy, integrity, and authenticity of the messages sent through it. This is commonly accomplished using cryptographic methods and protocols, which guarantee that only authorized recipients can view and interpret the transmitted information, safeguarding it from interception and modification.