Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

ammBoost: State Growth Control for AMMs

Nicolas Michel, Mohamed E. Najd, Ghada Almashaqbeh

TL;DR

ammBoost tackles AMM scalability by introducing a dependent sidechain architecture that offloads the bulk of AMM processing from the mainchain while keeping token custody and final state in TokenBank on the mainchain. It combines epoch-based deposits, pool snapshot-based trading, and TSQC-authenticated syncs within a PBFT-sidechain framework to preserve the AMM’s safety and liveness. Empirical evaluation on a Uniswap-inspired setup reports gas reductions of $96.05\%$ and chain-growth reductions of at least $93.42\%$, with throughput reaching $138.06$ tx/s and support for up to $500\times$ Uniswap’s daily traffic; it also achieves a $99.94\%$ reduction in transaction finality relative to Optimistic rollups. These findings indicate that ammBoost can substantially scale AMMs and potentially extend to other DeFi applications, while maintaining mainchain security and token custody guarantees.

Abstract

Automated market makers (AMMs) are a prime example of Web 3.0 applications. Their popularity and high trading activity led to serious scalability issues in terms of throughput and state size. In this paper, we address these challenges by utilizing a new sidechain architecture, building a system called ammBoost. ammBoost reduces the amount of on-chain transactions, boosts throughput, and supports blockchain pruning. We devise several techniques to enable layer 2 processing for AMMs, including a functionality-split and layer 2 traffic summarization paradigm, an epoch-based deposit mechanism, and pool snapshot-based and delayed token-payout trading. We also build a proof-of-concept for a Uniswap-inspired use case to empirically evaluate performance. Our experiments show that ammBoost decreases the gas cost by 96.05% and the chain growth by at least 93.42%, and that it can support up to 500x of the daily traffic volume of Uniswap. We also compare ammBoost to an Optimism-inspired solution showing a 99.94% reduction in transaction finality.

ammBoost: State Growth Control for AMMs

TL;DR

ammBoost tackles AMM scalability by introducing a dependent sidechain architecture that offloads the bulk of AMM processing from the mainchain while keeping token custody and final state in TokenBank on the mainchain. It combines epoch-based deposits, pool snapshot-based trading, and TSQC-authenticated syncs within a PBFT-sidechain framework to preserve the AMM’s safety and liveness. Empirical evaluation on a Uniswap-inspired setup reports gas reductions of and chain-growth reductions of at least , with throughput reaching tx/s and support for up to Uniswap’s daily traffic; it also achieves a reduction in transaction finality relative to Optimistic rollups. These findings indicate that ammBoost can substantially scale AMMs and potentially extend to other DeFi applications, while maintaining mainchain security and token custody guarantees.

Abstract

Automated market makers (AMMs) are a prime example of Web 3.0 applications. Their popularity and high trading activity led to serious scalability issues in terms of throughput and state size. In this paper, we address these challenges by utilizing a new sidechain architecture, building a system called ammBoost. ammBoost reduces the amount of on-chain transactions, boosts throughput, and supports blockchain pruning. We devise several techniques to enable layer 2 processing for AMMs, including a functionality-split and layer 2 traffic summarization paradigm, an epoch-based deposit mechanism, and pool snapshot-based and delayed token-payout trading. We also build a proof-of-concept for a Uniswap-inspired use case to empirically evaluate performance. Our experiments show that ammBoost decreases the gas cost by 96.05% and the chain growth by at least 93.42%, and that it can support up to 500x of the daily traffic volume of Uniswap. We also compare ammBoost to an Optimism-inspired solution showing a 99.94% reduction in transaction finality.

Paper Structure

This paper contains 23 sections, 3 theorems, 5 figures, 12 tables.

Table of Contents

  1. Introduction
  2. Our Contributions
  3. Background and Challenges
  4. Preliminaries
  5. System Design
  6. System Setup
  7. System Operation---Transaction Processing
  8. System Operation---Chain Management
  9. Security
  10. Implementation
  11. Performance Evaluation
  12. Experiment Setup
  13. Comparison with the Baseline
  14. Impact of Parameter Configuration
  15. Comparison with Rollups
  16. ...and 8 more sections

Key Result

Theorem 1

$\mathsf{ammBoost}$ preserves the safety and liveness of the underlying AMM.

Figures (5)

  • Figure 1: The $\mathsf{ammBoost}$ framework (Txs stands for transactions).
  • Figure 2: System setup.
  • Figure 3: $\mathsf{TokenBank}\xspace$ abstract functionality.
  • Figure 4: Summary rules ($\mathsf{userId}\xspace$ is the user ID and $\mathsf{posId}\xspace$ is the liquidity position ID). Updated liquidity pool balances are computed by TokenBank (as part of processing $\mathsf{Sync}\xspace$) based on the updated liquidity position and payout lists.
  • Figure 5: Gas cost and chain growth comparison.

Theorems & Definitions (8)

  • Remark 1
  • Remark 2
  • Theorem 1
  • Lemma 1
  • proof
  • Lemma 2
  • proof
  • Remark 3: On NFT-based liquidity positions