Why Zero Knowledge Rollups Need a New Consensus Mechanism

February 21, 2022

Why Zero Knowledge Rollups Need a New Consensus Mechanism

Zero knowledge rollups are widely seen as the end game for blockchain scaling because they can bundle thousands of transactions into compact proofs. But lunch refuses to be free and blistering throughput speeds come with a cost -- the dreaded threat of centralization.

What’s needed is a new way for blockchain participants to come to agreement on a single version of the truth, a consensus mechanism that doesn’t sacrifice performance while keeping potential cheaters at bay. The Polygon Hermez team, which is developing a zero knowledge implementation of the Ethereum Virtual Machine (EVM), has come up with a novel consensus mechanism: Proof-of-Efficiency (PoE). This is a concept for now, and it will keep evolving, as the team is exploring other options.

The PoE protocol aims to offer the following key properties:

Permissionless access to production of transactions batches on the rollup
Efficiency as key to network performance
Avoiding control by any single party
Protection from malicious attacks
Total validation effort proportional to the value in the network

Rollups exist on what is known as Layer 2 (L2), a collective term for scaling solutions that live outside the Ethereum mainnet, or Layer 1 (L1). Rollups still rely on L1 for security and data storage. This dual-layer topology is at the heart of how PoE works.

Sequencers and Aggregators

The PoE protocol for creating transaction batches consists of a two-step model that splits activities between different parties: Sequencers and Aggregators. Anyone can be a Sequencer or an Aggregator, these are permissionless roles.

Sequencers collect transactions from users on the rollup, then select and pre-process new batches of this Layer 2 data. Finally, they send transactions to Layer 1 to be recorded. Sequencers also deposit a fee in $MATIC token as an incentive for Aggregators to include the batch in a zero knowledge proof.

Aggregators are the parties that participate in the consensus protocol of PoE. The right to create the validity proof of a new state of the L2 is earned simply by being the first Aggregator to do it.

In this theoretical model, all transaction fees would be sent to the corresponding Sequencer of each batch, with a portion going to the Aggregator that included this batch into a validity proof. Fees deposited by Sequencers go to the Aggregators that included the respective batches into their validity proofs.

This mechanism avoids control by a single party and many of the potential attacks. But while any Sequencer can propose a batch, what happens when there is no participation of Aggregators? That situation is particularly likely at launch, before there is sufficient usage and the corresponding rewards for Aggregators. To address this issue, Polygon Hermez would launch a Boot Aggregator to make sure the network is fully operational during the bootstrapping phase.

PoE leverages the team’s experience with the existing Proof-of-Donation in Hermez 1.0. The goal of this preliminary concept is to share the ideas with the community and gather feedback and a more detailed discussion here. Polygon Hermez plans to have the first implementation of PoE in a testnet environment in the coming months.

Polygon’s ZK Thesis

Polygon is so bullish on the future of ZK, the core team made it a centerpiece of its strategic vision in the Zero Knowledge Thesis published in August. As part of that mission, Polygon committed $1 billion, a significant portion of the treasury, to our ZK-related efforts.

Over the past year, Polygon has emerged as a ZK powerhouse, assembling an unrivaled collection of R&D efforts under one roof. First was the merger with Hermez Network (now Polygon Hermez) in August. This was followed by the announcements of Polygon Nightfall, a privacy-focused rollup built in collaboration with EY, in September and Polygon Miden in November. In December, Polygon Zero came into being in a $400 million deal with pioneering ZK startup Mir.

We have made major progress toward making ZK technology a reality, but this is just the beginning. So tune in to our blog to keep up with the most exciting experiments in this space.

Let’s bring the world to Ethereum!