Research And Implementation Of Cloud Cache Service Construction Technology Based On NoSQL Database

In order to reduce the pressure on the relational database and increase the speed of data access of application,a caching mechanism can be established between the application and the database,but such caching mechanism usually works for a single service.To solve this problem,the mechanism can be applied to the cloud platform to provide services as a cloud cache.This thesis builds a cloud cache system based on the NoSQL database and distributed cache principle.Firstly,this thesis analyzes the related technologies and principles needed in the cloud cache,and determines the feasibility of building a cloud cache service.Through analyzing and comparing current caching software,some shortcomings and limitations of these software when running in cloud environment are pointed out.As a memory caching software,Memcached has better performance than Redis,but the functions of load balancing and redundant backup are not yet implemented.Also,mechanism of distribution is implemented in the client peer.Moreover,the two software do not provide multi-tenancy function,which means they can not meet demand of serving multi-application simultaneously.So,this thesis put forward some requirements,like the ability to serve multi-application simultaneously,high concurrency,scalability and high reliability.Thus,this thesis designs the architecture and partitions the modules of the entire cloud caching system,including load balancing module,network module,user module,storage module,and management module according to different functions.Secondly,this thesis designs and implements in detail for these modules of the cloud caching system.The load balancing module utilizes the CRUSH algorithm.By means of grouping calculation,it achieves a load balancing effect in both the distribution of request connections and the distribution of data placement.The network module defines the communication message format of the entire system and combines the asynchronous I/O network library to improve the system's ability of high concurrency.The user module provides user authentication to ensure system security.The storage module not only provides the function of data access,but also realizes the function of user data isolation by mapping logical data space to physical data space,thus meeting the requirement of serving multi-application simultaneously.Moreover,the storage module ensures data consistency between nodes through the mechanism of master-slave backup,which improves the system reliability.The management module can not only obtain the whole system running states through heartbeat mechanism,but ensure the system scalability.For the newly added nodes,the management module recalculates the distribution of data sets and completes the data migration.Finally,this thesis builds a cloud caching system based on NoSQL database.Detailed testing is performed and testing results are analyzed,which verifies that the cloud caching system meet the initial requirements and provide stable and efficient provisioning of caching services.