A Study And Optimization Of Cache Level In Collabrative Computing Platform

Redis is widely used for building a distributed cache system of high performance and large scale in cloud computing field. Redis cache system stores key-value pairs in memory to reduce the accessing pressure of traditional database, and improve the reading and writing speed. However in the pratical usage of the collabrative computing platform, Redis cache level will have very low throughput in a period, which bring negative affects to the speed and stability of the services. There is an urgent need to solve the problem.Firstly, we introduce the memory, persistence, event processing, cluster model and the principle of Redis, make a full comparision of Redis and Memcached to illustrate the reason why we use Redis to build cache level.Then, we analyse why the low throughput appears in Redis cache level. We narrow the research fields by the problem's necessary conditions. We check and analyse the data structure and the algorithm flow related with the problem to locate the problem. We propose a new improvement to solve the problem, so that we can have a better cache level in collabrative computing platform which can provide quick and stable services.Finally, we conduct an experiment of comparison which shows that our improvement can solve the low throughput problem we meet. The new Redis cache level works better in our collabrative computing platform.