DESIGN OF DATABASE STORAGE HIERARCHY AND RELIABILITY MECHANISM IN CLOUD-NATIVE ENVIRONMENT

Abstract

With the in-depth popularization of cloud computing and microservice architecture, cloud-native databases have become the core infrastructure supporting the storage of massive business data and high-concurrency access. The tightly coupled architecture of traditional databases suffers from insufficient resource elasticity, high storage costs, and weak fault tolerance, making it difficult to adapt to the dynamic scaling, on-demand billing, and high-availability operation requirements of cloud-native environments. To address the above pain points, this paper proposes a database storage hierarchical architecture for cloud-native environments. Combining the design concept of compute-storage separation, a three-tier storage system consisting of a hot data cache layer, an online business storage layer, and a warm-cold data archive layer is constructed. Meanwhile, a multi-dimensional reliability guarantee mechanism is designed, covering four core modules: data consistency, fault self-healing, disaster recovery and backup, and anomaly isolation. An experimental environment was built for performance verification based on the TPC-C and TPC-H standard test sets. The results show that this architecture can effectively improve the storage resource utilization rate by more than 35%, reduce the cold data storage cost by 60%, and control the fault recovery time within 10 seconds, meeting the high-performance, low-cost and high-reliability storage demands of databases in cloud-native scenarios.