Key points are not available for this paper at this time.
While sharding technology helps solve performance bottlenecks in traditional blockchain networks, it also introduces new challenges. Random node allocation may lead to uneven distribution of malicious nodes, causing performance differences and security risks. Existing reputation-based sharding schemes often overlook node performance characteristics and fail to address security concerns related to leader election. In addition, full sharding reorganisation drastically reduces the performance of the entire blockchain, both in terms of overhead and time. In this paper, we propose a reputation-based sharding dynamic reorganisation network sharding scheme, which integrates node performance and behavioural characteristics into the reputation computation, and at the same time adds alternative leaders to maintain the stability of the sharding system. Based on this, through the partial reorganisation algorithm of the sharding, the nodes are reasonably allocated to balance the computing power of each shard, effectively stimulating the liveness of the sharding system and improving overall safety and performance. Simulation results show that this sharding scheme effectively reduces the consensus failure probability, solves the collusion attack problem caused by the aggregation of malicious nodes within shards, and reduces the latency of the blockchain system. This provides strong support for the reliability and performance of the blockchain sharding system.
Zhang et al. (Tue,) studied this question.