GPChain: Optimizing Cross-Shard Transactions and Load Imbalance in Sharded Blockchain Networks

Sharded blockchain provides a better linear scalability and scalability compared to traditional blockchain by dividing the blockchain into several disjoint shards, making it a more suitable blockchain solution for the Internet of Things (IoT). However, the transaction data generated by devices in the network also follows a power-law distribution, resulting in a large number of cross-shard transactions and shard load imbalance due to popular addresses when partitioning transactions into shards. Moreover, the atomicity of transactions requires cross-shard transactions to wait for confirmation from multiple shards, further prolonging the confirmation time due to transaction congestion in shards. In this paper, we propose GPChain (Graph Partitioning Chain), a graph-based sharding blockchain transaction optimization scheme. We address the challenges of load imbalance in shards caused by power-law distribution and excessive cross-shard transactions through an enhanced graph partitioning algorithm. Experimental results demonstrate that GPChain outperforms other solutions in terms of cross-shard transaction ratio and shard load balancing.