Research On Adaptive Caching Mechanism Of Relational Cloud Database

The combination of traditional database and cloud computing technology has spawned a cloud database,which provides virtualization of database resources,with on-demand use,scalability,and high availability.Its importance and value are increasingly apparent.In the Internet scenario,the relational cloud database faces variable data access patterns and load pressures,resulting in low efficiency of existing cache management and large fluctuations in system performance.Therefore,enhancing the adaptive ability of the cache subsystem is of great significance for improving the overall stability and throughput of the system.The relational cloud database adaptive cache framework is designed by using the storage computing separation technology.In this framework,the cache partition classification management strategy is designed.By dividing the cache space into multiple logical partitions and mapping the data pages to the corresponding partitions according to the data access mode and the data page's own attributes,the cache concurrent access conflicts are alleviated.Based on this strategy,the data page adaptive persistence strategy is designed to dynamically coordinate persistent worker threads and optimize disk write efficiency.According to the designed strategy,the dynamic feedback cache replacement algorithm and the cache capacity elastic scaling algorithm are designed and implemented to manage and reclaim the data pages in the cache to improve the cache hit rate and utilization.Design a cache object preloading strategy and implement a cache load stress evaluation algorithm.Improve the out-of-service service timeliness by preloading hotspot data into the cache.Evaluate the cache read overload node,map its hot cache status to the newly added node cache,and balance the read load pressure of each node.By establishing a relational cloud database test environment,and using Sysbench test tool to simulate the data access mode under different load scenarios,the related strategies and algorithms are designed and implemented to verify and analyze the functionality and effectiveness of the adaptive cache framework.