Distributed key-value stores are widely adopted to support elastic big data applications, leveraging purpose-built consensus algorithms like Raft to ensure data consistency. However, through systematic analysis, we reveal a critical performance issue in such consistent stores, i.e., overlapping persistence operations between consensus protocols and underlying storage engines result in significant I/O overhead. To address this issue, we present Nezha, a prototype distributed storage system that innovatively integrates key-value separation with Raft to provide scalable throughput in a strong consistency guarantee. Nezha redesigns the persistence strategy at the operation level and incorporates leveled garbage collection, significantly improving read and write performance while preserving Raft's safety properties. Experimental results demonstrate that, on average, Nezha achieves throughput improvements of 460.2%, 12.5%, and 72.6% for put, get, and scan operations, respectively.

In multi-domain deployment scenarios, leader-based distributed systems face high latency challenges when processing cross-domain requests. To address this challenge, we propose a novel, network-topology-aware dynamic leader scheduling mechanism. This mechanism utilizes a system model to dynamically evaluate the system latency when the leader is located in different domains, incorporating real-time state information such as network latency, node distribution, and geographic workload distribution. Based on the evaluation results, the mechanism schedules the leader to the optimal domain to minimize latency. Experimental results indicate that it can effectively reduce system latency and improve system Quality of Service (QoS).