The New CAP Theorem on Blockchain Consensus Systems

A theoretical limit to decentralization: Consensus, Autonomy, and Performance cannot be simultaneously optimized

A. Anagnostakis; E. Glavas· Future Internet· 2025· DOI 10.3390/fi17040157

The core problem

Inspired by Eric Brewer's CAP theorem for distributed databases, which states that Consistency, Availability, and Partition tolerance cannot all be guaranteed simultaneously, this work introduces a new CAP theorem for autonomous consensus systems. The authors propose that in blockchain consensus systems, at most two of three elementary properties—Consensus achievement (C), Autonomy (A), and entropic Performance (P)—can be optimized simultaneously in the generic case. This imposes a theoretical limit on decentralization, affecting scalability, security, and real-world adoption. The study aims to formalize and analyze this tradeoff using the IoT micro-Blockchain as a universal, minimal, consensus-enabling framework.

The formal proof demonstrates that for a homogeneous system, the three properties cannot be concurrently optimized. This means that any attempt to maximize all three simultaneously is impossible in the generic case. Empirical benchmarking of Bitcoin, Ethereum, and Hyperledger Fabric supports the theoretical findings, showing that these systems exhibit tradeoffs consistent with the theorem. For instance, Bitcoin prioritizes consensus and autonomy at the expense of performance, while Hyperledger Fabric may prioritize performance and consensus with reduced autonomy. The results highlight an intrinsic limitation on the design and optimization of distributed blockchain consensus mechanisms.

Inspired by Eric Brewer's CAP theorem for distributed databases, which states that Consistency, Availability, and Partition tolerance cannot all be guaranteed simultaneously, this work introduces a new CAP theorem for autonomous consensus systems. The authors propose that in blockchain consensus systems, at most two of three elementary properties—Consensus achievement (C), Autonomy (A), and entropic Performance (P)—can be optimized simultaneously in the generic case. This imposes a theoretical limit on decentralization, affecting scalability, security, and real-world adoption. The study aims to formalize and analyze this tradeoff using the IoT micro-Blockchain as a universal, minimal, consensus-enabling framework.

The authors utilize the IoT micro-Blockchain framework to model consensus systems. They define quantitative functions relating each property (C, A, P) to the number of event witnesses in the system. The framework allows for formal analysis of mutual exclusions among the properties. For a homogeneous system, they formally prove that (A), (C), and (P) cannot be optimized simultaneously. The proof is constructed using mathematical formalism, and the findings are validated through empirical data benchmarking of large-scale blockchain systems: Bitcoin, Ethereum, and Hyperledger Fabric. The methodology combines theoretical proof with empirical validation to demonstrate the inherent limitations.

The new CAP theorem provides a theoretical foundation for understanding the limitations of blockchain systems. It suggests that decentralization (autonomy) cannot be maximized alongside consensus and performance, impacting scalability and security. This has implications for real-world adoption, as designers must make explicit tradeoffs. The use of the IoT micro-Blockchain framework offers a minimal model that can be extended to other consensus systems. The authors conclude that concurrent optimization of C, A, and P is unattainable, revealing an inherent constraint in distributed consensus. Future work may explore relaxing assumptions or alternative frameworks to mitigate these tradeoffs.

CS practitioners and researchers

Opening member content…