Skip to main content

Design Goals

We draw on the theoretical goals outlined in "Strategyproof Computing", a paper that illustrates how to think about incentives in the design of distributed systems: 

1. Incentives-first
2. Utility-based
3. Simple
4. Open
5. Decentralized

and the practical ones which led to the implementation of Serenity:

1. Simplicity
2. Long-term stability
3. Sufficiency
4. Defense in depth
5. Full light-client verifiability

I.

Permissionless

It should be possible for anyone to deploy any kind of builder solidity contract. This is inspired by the "open" principle from Strategyproof computing, which states: "Our belief is that an open marketplace will naturally lead to mechanisms with the 'right scope' and 'right complexity'."

We cannot preempt every kind of MEV that will exist, or the range of methods which can be concocted to extract and distribute it. Therefore, we aim to enable anyone, anywhere to create and deploy novel contracts in response to unpredictable needs. The public nature of these permissionless contracts should help the entire ecosystem remain proactive, as permissionless access should imply a contestable marketplace of designs.

II.

Simple

Strategyproof computing argues that we ought to "simplify the decisions facing participants in distributed multi-agent systems", and Vitalik points out that simplicity matters "because it (i) minimizes development costs, (ii) reduces risk of unforeseen security issues, and (iii) allows protocol designers to more easily convince users that parameter choices are legitimate".

Simplicity implies flexibility with respect to dependencies, such that we can upgrade from SGX to MPC, or deploy better cryptographic solutions like FHE, or introduce some maximally robust and sufficiently decentralized combination of the above.

III.

Sufficient

This is inspired by the Serenity design rationale, which states that it should be "possible to build as many classes of applications as possible on top of the protocol".

We strive for the same generality with SUAVE: it should be possible to use SUAVE to create any kind of application to do with order flow, block building, commitments, instantaneous privacy, and any other MEV application we have yet to imagine.

IV.

Private

The operation of mechanisms for ordering information and distributing value must be public and verifiable. The data which passes through those mechanisms must not be. Privacy is a human right and, in the context of MEV, it enables us to create more economically efficient systems that are resilient to centralization of wealth and power over time, and to malicious manipulation in any given moment.

We must provide people with the ability to keep their data hidden and only reveal it to those whom they trust or wish to interact with.

V.

Decentralized

Allocating resources without a central authority is hard because there is a trade-off between revenue optimality and efficiency. People should benefit maximally from their information, and finding or correcting inefficiencies should carry sufficient reward.

SUAVE is simultaneously a search for the most efficient mechanisms for value extraction, and the most optimal mechanisms for value distribution. How such mechanisms work cannot be invented in isolation, it must be discovered together (like the price in an auction). No-one should have undue power in this process of discovery. SUAVE unifies insights from information theory, computer science, economics, and sociopolitics to ensure such power remains properly decentralized.

VI.

Credible

This is inspired by Strategyproof computing, which states that, "the self-interest of participants in distributed systems should be explicitly addressed by system designers."

It should be possible to quantify who benefits, and how much they benefit relative to the average.

We do not mean "who" in the personal sense, but rather in terms of the type of actors involved. Specifically, users should enjoy the maximum value possible while ensuring SUAVE is adopted by all parties, with the rest of that value cascading down to those doing the most relevant work.