A cloud service-oriented distributed database intelligent library and table splitting method and system

Abstract

This invention discloses a method and system for intelligent sharding and partitioning of distributed databases for cloud services, relating to the fields of cloud computing and distributed database technology. The method includes: real-time SQL parsing and instance load collection; obtaining future popularity distribution using a hybrid prediction model, while simultaneously constructing a data affinity graph of tables and fields; automatically generating candidate sharding strategies through multi-objective optimization, aiming to minimize cross-shard queries, load skew, and data migration volume; automatically executing the strategies after cost-benefit evaluation; and completing data migration and route switching online based on canary routing and cloud-native incremental synchronization. This invention achieves intelligent generation, proactive balancing, and seamless business switching of sharding strategies, significantly reducing the complexity of distributed transactions and operations, and deeply collaborating with cloud platform elasticity to continuously maintain optimal data distribution.