AASC: Activity-Adjusted Stake Consensus for Scalable IoT Blockchains

A lightweight, ML-augmented consensus protocol for dynamic IoT networks

Susmita Mondal; Adarsh Shrivastava; Jatin Lohar; Suchetana Chakraborty· International Conference of Distributed Computing and Networking· 2026· DOI 10.1145/3772290.3772307

The core problem

The Internet of Things (IoT) has rapidly expanded, enabling decentralized decision-making and data exchange. Blockchain technology offers a promising solution for transparent, trustless, and traceable data management in IoT without centralized infrastructure. However, traditional consensus mechanisms like Proof of Work (PoW) and Proof of Stake (PoS) are ill-suited for resource-constrained IoT devices due to high energy consumption, computational overhead, and vulnerability to attacks in dynamic environments. This paper proposes Activity-Adjusted Stake Consensus (AASC), a lightweight, iterative, and machine learning-augmented consensus protocol designed specifically for Blockchain-enabled IoT. AASC continuously updates a reputation score (Consensus Score) for each node based on historical and real-time behavior, enabling adaptive consensus in heterogeneous and ever-changing networks. The key contributions include: (1) a novel consensus mechanism that transcends rule-based approaches, (2) linear O (n) communication and O (n lo g n) computational complexity per node, and (3) resilience against Sybil, collusion, and monopolization attacks through randomized validator selection and swift p

Innovation

Extensive experiments were conducted to evaluate AASC against leading lightweight consensus protocols (e.g., PoS variants, PBFT) in terms of fairness, security, and scalability. Key findings include: - **Fairness**: AASC achieves a Gini coefficient of 0.12 for validator selection, indicating high fairness, compared to 0.35 for PoS. - **Security**: Under a Sybil attack with 30% malicious nodes, AASC maintains a consensus success rate of 95%, while PoS drops to 70%. Collusion attacks are mitigated by randomized selection, reducing the probability of collusion by 80%. - **Scalability**: AASC scales linearly with the number of nodes; for 1000 nodes, the average consensus latency is 2.3 seconds, and communication overhead is 1.2 MB per round, significantly lower than PBFT (5.6 seconds, 4.8 MB). - **Overhead**: The ML-based score update adds negligible overhead (0.5 ms per node per round), making it suitable for resource-constrained devices. These results demonstrate AASC's superiority in dynamic IoT environments.

The Internet of Things (IoT) has rapidly expanded, enabling decentralized decision-making and data exchange. Blockchain technology offers a promising solution for transparent, trustless, and traceable data management in IoT without centralized infrastructure. However, traditional consensus mechanisms like Proof of Work (PoW) and Proof of Stake (PoS) are ill-suited for resource-constrained IoT devices due to high energy consumption, computational overhead, and vulnerability to attacks in dynamic environments. This paper proposes Activity-Adjusted Stake Consensus (AASC), a lightweight, iterative, and machine learning-augmented consensus protocol designed specifically for Blockchain-enabled IoT. AASC continuously updates a reputation score (Consensus Score) for each node based on historical and real-time behavior, enabling adaptive consensus in heterogeneous and ever-changing networks. The key contributions include: (1) a novel consensus mechanism that transcends rule-based approaches, (2) linear O (n) communication and O (n lo g n) computational complexity per node, and (3) resilience against Sybil, collusion, and monopolization attacks through randomized validator selection and swift penalization.

AASC operates in iterative rounds, where each IoT node maintains a Consensus Score S_{i} that reflects its activity and trustworthiness. The score is updated using a machine learning model that considers factors such as participation frequency, data accuracy, and responsiveness. The consensus process involves the following steps:

Why it matters

AASC addresses critical challenges in IoT blockchain consensus by integrating activity-based reputation with machine learning. The use of a Consensus Score allows for adaptive trust management, ensuring that active and honest nodes are rewarded while malicious ones are penalized. The randomized validator selection prevents monopolization and collusion, as attackers cannot predict or control the selection process. The linear communication complexity makes AASC scalable for large IoT networks, and the O (n lo g n) computational complexity per node is feasible for devices with limited resources. However, the reliance on an ML model introduces potential vulnerabilities, such as adversarial attacks on the model itself. Future work should explore robust ML techniques and formal verification of the protocol. Overall, AASC paves the way for practical blockchain adoption in IoT by balancing efficiency, security, and fairness.

Who should read this

CS practitioners and researchers

Opening member content…