B+Avl Trees: Towards Data Structures for Robust and Efficient Blockchain State Synchronization

This paper addresses state transfer in blockchain systems, crucial for new or recovering peers. Current systems use periodic snapshots and cryptographic structures like Merkle trees for efficient validation in Byzantine fault tolerant scenarios. Recent improvements use chunk-based data structures (e.g., AVL* trees, Merkle B+trees) to reduce overhead and enable efficient chunk-level validation. This paper introduces B+AVL trees, a novel chunk-based structure that significantly advances the state transfer process. B+AVL trees combine the balancing mechanism of AVL trees with Merkle hashing for validation, enabling efficient and secure state transfer. They are more spaceefficient and simpler to manage than AVL* trees, and they use more compact validation proofs than Merkle B+trees. B+AVL trees represent a major improvement by balancing binary trees and managing chunks in a way that optimizes performance and data validation, ensuring robust state synchronization in blockchain systems. The paper thoroughly assesses B+AVL trees and compares them to competing approaches.