Honeybee: Byzantine Tolerant Decentralized Peer Sampling with Verifiable Random Walks

TL;DR

Honeybee tackles the challenge of secure, scalable, decentralized peer sampling in permissionless blockchains under Byzantine/Sybil threats. It introduces verifiable random walks (VRW) combined with table consistency checks (TCC) to achieve near-uniform sampling and detect equivocation, implemented through address tables with bilateral peering and expiry. The approach demonstrates strong resilience against a wide range of adversarial strategies, outperforming Kademlia and GossipSub in simulations and achieving $\epsilon$-uniform sampling with $\epsilon=0.03$ across substantial fractions of adversarial nodes. Practical deployment considerations, including secure randomness sources and overheads, are discussed, with implications for data availability sampling and sharding in blockchain networks.

Abstract

Popular blockchains today have hundreds of thousands of nodes and need to be able to support sophisticated scaling solutions$\unicode{x2013}$such as sharding, data availability sampling, and layer-2 methods. Designing secure and efficient peer-to-peer (p2p) networking protocols at these scales to support the tight demands of the upper layer crypto-economic primitives is a highly non-trivial endeavor. We identify decentralized, uniform random sampling of nodes as a fundamental capability necessary for building robust p2p networks in emerging blockchain networks. Sampling algorithms used in practice today (primarily for address discovery) rely on either distributed hash tables (e.g., Kademlia) or sharing addresses with neighbors (e.g., GossipSub), and are not secure in a Sybil setting. We present Honeybee, a decentralized algorithm for sampling nodes that uses verifiable random walks and table consistency checks. Honeybee is secure against attacks even in the presence of an overwhelming number of Byzantine nodes (e.g., $\geq50\%$ of the network). We evaluate Honeybee through experiments and show that the quality of sampling achieved by Honeybee is significantly better compared to the state-of-the-art. Our proposed algorithm has implications for network design in both full nodes and light nodes.

Figures (11)

• Figure 1: Honeybee illustrations: (a) an example of Honeybee verifiable random walk (VRW) sampling with three hops; (b) how VRW works from hop to hop; (c) how table consistency check (TCC) works among nodes.
• Figure 2: In a Honeybee network with 50% dishonest nodes, all dishonest nodes cluster to attack a victim whose initial address table contains only one honest outgoing neighbor (an unlikely adverse scenario). We conduct five separate such experiments and present the worst-case outcome for the victim.
• Figure 3: Nodes adhere to protocol: Sampling distribution from a random observation node in 100,000 epochs. Every node follows its protocol. Node IDs are sorted by frequency in ascending order. True uniform sampling distribution is shown as a dashed line.
• Figure 4: Quality of sampling when nodes adhere to protocol: (a) displays comparisons of Honeybee, GossipSub, and Kademlia with the true uniform sampling distribution; (b) displays the Chi-Square test results with p-value of 0.05 shown as a dashed line; (c) displays the Chi-Square test results when 30% of the nodes are idle.
• Figure 5: An honest node attacked by 5%, 10%, and 20% of dishonest nodes: Single random honest node under attack in Honeybee, GossipSub, and Kademlia.

Theorems & Definitions (7)

• lemma 1
• theorem 1
• theorem 2
• theorem 3
• lemma 2
• lemma 3
• lemma 4