High-Level Architecture Overview

The architecture of the Arcium Network is designed to be exceptionally versatile and flexible, making it suitable for all possible distributed confidential computing needs. Confidential computing tasks are broken down into individual Computation definitions which are executed within dedicated MPC environments called MXEs. These MXEs are operated on top of Clusters which are collections of Arx (MPC) nodes. In classical computing terms, Clusters can be thought of as analogous to the hardware of the system, MXEs represent the system's state, Computation definitions act as functions, and each scheduled Computation resembles an individual function call instance.

The Arcium Network uses programs running on the Solana blockchain to manage and orchestrate all aspects of the Network's state and services. This includes an on-chain mempool architecture that houses all pending Computations awaiting execution by the Network. Additionally, participant incentivization is also handled on-chain, with compensation for confidential computing executions flowing directly from Computation Customers to Arx Operators and 3rd-Party Delegators.