A Greedy Account Relocation Algorithm With Load Minimization for Sharded Blockchains

As blockchain technology is adopted in innovative fields like the Internet of Things, software-defined IoT, QoS-aware software-defined IoT, software-defined Internet of Multimedia Things, and streaming in software-defined Internet of Vehicles, it faces scalability challenges, including rapidly increasing transaction throughput, confirmation delays, and rising fees. An emerging technique, sharding is a solution to scale the blockchain by processing transactions in parallel. The sharding splits the main network into subnetworks, called shards, to process transactions. Sharding techniques for blockchain scalability face challenges in minimizing cross-shard transactions (CSTs) and avoiding unnecessary or delayed account relocations due to random shard assignments. In this article, we introduce a greedy account relocation algorithm to reduce the number of CSTs. In our algorithm, the accounts are first sorted according to the number of CSTs they are responsible for, and then the accounts are reassigned to a target shard if the relocation satisfies given constraints. Our algorithm can reduce the CSTs more than 30% compared to the existing relocation algorithms, still keeping load balancing between shards.