The Anoma Protocol and its Design Process

Fri, Sep 15, 2023, 01:00 PM · 22:20

This talk will give an introduction to the Anoma protocol architecture and its design process, to share what we learned and elicit feedback on how we could improve both. Anoma is a distributed operating system for intent-centric counterparty discovery and privacy-preserving computation on linear and non-linear resources with heterogenous trust assumptions. The protocol architecture has the following (non-exhaustive) goals and design process behind it: Goal: Identify good component boundaries, to unbundle features and improve modularity. Process: - Analyze the outcomes of protocols that came before us, learn from their successes and problems. - Example components: Consensus Algorithms, Network Transport, Programming Environment, State Machines Goal: Maximize composability of components. Process: - Try to bridge the gap between real world systems and models from Programming Language Theory, Automata Theory and Category Theory. - Come up with implementation agnostic interfaces using above theory. Goal: Maximize flexibility of the protocol stack, since changes in a deployed system are costly, proportional to the significance of the change. Process: - Decouple dimensions of choice as much as possible, to enable choices to be made independently of each other. - Minimize the amount of design decisions and make them at the least-significant place. - Wherever possible, try to push decisions from "design-time" to "runtime", so users can make them as circumstances change. We will give a high level overview of the architecture and explain, using examples from it, how they are motivated by the process and how they move us closer to the above goals.