Investigating Impacts of Uneven Node Distribution and Cross-Shard Transactions on Sharding IoT-Blockchain Systems

Abstract

The scalability limitations of blockchain hinder its application in large-scale IoT-Blockchain systems. Sharding improves scalability by dividing the IoT-Blockchain network into multiple shards. However, most of the previous studies assumed a fixed node distribution within shards, overlooking the dynamic nature of real-world IoT-Blockchain systems, hence it may lead to inaccurate performance evaluations. Additionally, cross-shard transactions significantly impact system performance. The Broker role, introduced by BrokerChain, a blockchain sharding protocol that utilizes broker accounts, has been proposed to mitigate this issue. However, existing research lacks an in-depth analysis of the number of brokers (BrokerNum) and its interaction with core sharding parameters, such as the number of shards (ShardNum) and the number of nodes per shard (NodesInShard). In this paper, we emulated a sharding IoT-Blockchain system using the BlockEmulator platform and introduced random uneven node distribution scenarios to analyze their impact on key performance metrics, including transactions per second (TPS), transaction confirmation latency (TCL), and the cross-shard transaction ratio (CTX). We also systematically investigated the influence of BrokerNum, ShardNum, and NodesInShard on system performance. To improve efficiency, we developed an automated experimental process that significantly reduces execution time. The results show that variations in node distribution have a significant impact on system performance, emphasizing the importance of an adaptive sharding strategy. Moreover, the relationship between BrokerNum, ShardNum, and NodesInShard introduces a notable performance trade-off. Properly tuning these parameters is crucial for enhancing overall system efficiency.