Research And Application Of Optimization Technique Based On Redis For Information Storage

Information storage technology is an important part of digital technology. The explosive growth of digital information, have been set higher demand on information storage, and information storage systems should improve their performance substantially. So far, the methods to improve the storage performance mainly include storage virtualization, storing in memory, interoperability with multiple platforms and data sharing, etc. From a technical perspective, the performance of the present storage systems, cannot meet the demand of high concurrency. This article choose’Redis’, a new storage technique, aimed at the high concurrency of web applications, to optimize the information storage technology based on Redis to meet the demand of high concurrency. Firstly, for particular cases that the reading and writing ratio is about10:1in internet applications, design a new information storage scheme based on Redis, combed with relational database to realize separation of heterogeneous reading and writing. Second, reasonable optimization is made to Redis, to further improve its performance.For heterogeneous data synchronization, make use of DML in relational data to access the data to achieve full amount of heterogeneous migration. To realize real-time synchronization, make use of master-slave replication mechanism of relational database, pretending to be slave database to support delta copy. It supports converting various traditional relational database to data structure of Redis, have very strong commonablility. Optimizing Redis in three aspects, included choosing memory allocator, data persistence and network delay. Putting forward load-sharing algorithm based on weighted consistent hashing to balance loading pressure.In design and implementation, we set up relevant data-serving platform to realize conversion and synchronization among heterogeneous data. By experiment, verifying conversion and real-time synchronization among heterogeneous data in the data-serving platform, and validating the high performance through high concurrency test. The results show that, heterogeneous data can transform and be synchronized through data-serving platform. In high concurrency cases, performance of internet application based on data-serving platform is better of that based on pure relational database.