On the Trade-Off between Degree of Decentralization and Scalability in Proof-of-Stake–based Blockchain Systems

Abstract

Proof-of-stake-based (PoS-based) blockchain systems exhibit a trade-off between degree of decentralization (DoD) and scalability: equitable participation of validating nodes can raise latency and degrade throughput, while scaling can reduce equitable participation. To meet application requirements, software architects must balance these dimensions using comparable metrics. We curated a compact metric set for DoD (block-proposal randomness [Shannon entropy], token concentration [HHI], wealth distribution [Gini]) and for scalability (transaction throughput, first-inclusion latency) and applied it to Tezos Hangzhou 2 (Octez v12.3). Across 1,573 controlled configurations with varying validating-node count n, block interval, and block size limit, we normalized DoD by n to enable cross-configuration comparison. Under this normalization, DoDscalability associations were weaker and not uniformly negative. Scalability was primarily influenced by validating-node count. Our main contributions are a benchmarking approach with curated metrics, an empirical mapping of configuration levers to both dimensions, evidence for the dominant role of n, and guidance for configuring PoS-based blockchain systems.