Research On Security Extension And Performance Optimization For Distributed Cache

With the rapid development of network technology,the data are characterized by large-scale and high-concurrency,the traditional relational database has been unable to meet their real-time,fast response and other needs in the face of these circumstances.Because the relational database data stored in the disk,I / O bottlenecks constrain data access speed,In addition,it take too many functions,such as isolation levels.Redis which is a kind of high-speed cache-based database,its data model is based on Key-Value.It has open source,networked,memory-based,data persistence and other features.It is not bound by the characteristics of the relational database,such as paradigm.Redis can effectively improve the system real-time and fast response efficiency as a relational database cache,Redis largely solves the problems faced by relational databases,but it has some deficiencies in some areas,such as security and load balancing.In this thesis,we will design a security extension mechanism which is suitable for Redis through the current mainstream information security encryption and decryption technology analysis.We will compare various encryption algorithms from performance and security,we also encrypt the data hierarchically and take into account the efficiency issues.We will achieve secure access to the data and access control.In addition,we can guarantee the confidentiality of data during transmission,storage security.We will implement the Redis cluster load balancing.These will let Redis have a better application by further improving the Redis function,the specific content of this paper has three aspects.We analyze Redis source code.We are familiar with and master the data structure,program running process,Redis cluster and some other Redis internal operating mechanism by reading Redis source code and some relevant information.The goal of these tasks is to modify the source code to implement the corresponding function.We analyze Redis' s security flaws and achieve a security extension to Redis.We build the security level of the data and implement role-based access control for Redis.We protect the data and key by encryption Algorithm.We implement the load distribution of the Redis cluster under isomorphism,achieve the load distribution under heterogeneous.We also monitor real-time usage of node memory in the Redis cluster.Redis will transfer load for this node when the node memory capacity is small.This function can reduce the load for the node and let Redis cluster`s memory resources be used reasonably.