The Design And Implementation Of Distributed Redis High Availability Cluster

As a highly efficient Key-Value database storage system,Redis supports multiple types of storage methods and data operations.In addition,its protocol type is simple and widely used in data storage and caching.In the case of a large data volume,it is necessary to deploy multiple Redis instances to solve the memory bottleneck of a single physical machine.Currently,the official Redis Cluster solution can deploy multiple Redis instances,but it's too complex to maintenance and upgrade that is not suitable for production.This thesis analyzes the requirements of Redis cluster under the scenarios of large data and high concurrent business.The thesis compares some popular Redis distributed clusters products and then applies proxy mode into data fragmentation.A high available Redis cluster is designed and implemented based on the characteristics of Redis.The Proxy layer is designed in this scheme,and it provides a unified calling interface while shields the specific details of the underlying Redis cluster.The Proxy can implement high availability by deploying multiple instances and load balancing,and maintain data consistency with ZooKeeper which can store routing information as the configuration center of these instances.This thesis mainly describes the design and implementation of Proxy layer.The Proxy layer is divided into four modules:Net 10,Proxy s network framework;Packet Handler,analyse and process Redis communication protocol;ZKRouter module,interact between the Proxy and the ZooKeeper and initialize the configuration information such as the proxy routing;The Migrate module,help the Proxy to provide services while the external data is during migration.The clustering scheme proposed in this paper has been put into use.It shows that the scheme meets the requirements of high throughput and high availability,thus the correctness and rationality of the scheme is verified.