The rapid growth of Internet of Things (IoT) devices is creating enormous amounts of data in large-scale IoT systems. In such systems, the efficiency, trustworthiness, confidentiality, and integrity of data storage and retrieval are significant challenges. In this paper, we propose a dual-blockchain distributed data structure as a highly scalable solution for resource-constrained IoT systems. Our proposed model is based on a distributed hash table (DHT) which is implemented and benchmarked in an experimental IoT network environment. Performance analysis and experimental results highlight that our dual-chain model provides better scalability in terms of space consumption, latency and computational complexity compared to baseline single-chain models. Security analyses are provided to discuss the feasibility of distributed trust management, fault tolerance, as well as confidentiality of data stored in our model and conflict avoidance mechanisms.