分布式环境下的频繁数据缓存策略

FREQUENT DATA CACHING STRATEGIES IN DISTRIBUTED ENVIRONMENT

大数据环境下利用分布式缓存技术能够提供高性能、高可用的数据查询。针对轻量级数据库应用的频繁数据缓存策略具有高效、易扩展的优点,更有利于轻型分布式数据库应用的查询优化改进。因此,通过分析用户行为和用户查询特征,研究针对近期频繁查询数据的数据缓存策略,能够预测高命中率的缓存数据,提高数据查询效率。首先分析并给出查询频繁度的定义,其次根据时间因素对缓存数据选取的影响细化用户查询操作,并通过查询数据的查询频繁度应对查询过程中不同的缓存命中情况整合节点间的缓存数据。最后,实验证明该数据缓存策略具有较高的数据命中率,能够提高数据查询的效率。实现方面可根据实际需要采用不同的缓存属性组合,具有良好的易扩展性。

Using distributed caching technology can provide high performance and high availability data query in large data environment. The frequent data caching strategies for lightweight applications have the advantages of high efficiency and easy extension. Especially, it is beneficial to improve the query optimization for lightweight distributed database system. Therefore, this research studies the data caching strategies for the recent frequent query data by analyzing the characteristics of user behavior and user query. It can predict the high hit rate of the cache data and improve the efficiency of data query. Firstly, the definition of the frequency of the query was analyzed and given. Secondly, we refined the operation of the user s query according to the influence of the time factor for the cache data. And we dealt with the different cache hits in query process through the data query frequency. Then, we integrated the cached data between nodes. Finally, the experimental results showed that the data caching strategy has a high data hit rate. It also can improve the efficiency of data query. According to the actual needs, the implementation can use different combination of cache attributes, and possesses a good scalability.