The Study And Implementation Of Data Cache Based On MEMCACHED

With the rapid growth of WEB applications and the amount of data,the response time of the page and the stability of the system becomes an important indicator of the user experience.The most important factor affects the system response time and system stability is the query performance of the data.In order to speed the data query process,the application of Memcached caching system is used to cache the data with frequent query.Different data caching schemes are proposed for different data and application scene.Because of data is cached in the server's memory,and the client will read data directly from memory when the client requests data.Due to the limitation of server's memory,one server can't cache massive data and it can't deal with high concurrent requests.A distributed solution is proposed.With this distributed deployment scheme,the paper analyze the problems that will appear in the process of using different data dispersing algorithms,and study the specific implementation of adding & deleting server nodes in distributed Memcached system.The consistent hashing algorithm and virtual node concept applied to the distributed architecture of Memcached,it can improve the performance of the cache system when one server is down.Binary search algorithm is used when find the server node in consistency hash algorithm,it make the lookup faster.Due to the data in memory will be lost,and cannot be recovered when the server is restarted or the server is down.And its mechanism of LRU will clear the data which is not expired.In view of the features that the data of Memcached cache server is easily lost,this paper proposes a solution to the persistence of cached data.It solve the problem that the cache data is easily lost.When one of the cache server down,it can also add a new replacement server,and the cached data in old server will be recovered from persistent data.The research results in this paper can be applied to web cache system with high concurrent request or with large amount of data,which solves many problems such as data loss,insufficient memory and performance bottleneck.According to the experimental results,the speed of data access is increased several times,and the cache server can be extended indefinitely.The distributed solution and data persistence scheme proposed in this paper have some reference for other caching systems.