Design And Implementation Of The ROLAP Cache Mechanism Based On The Incremental Learning Naive Bayesian

In the era of big data, due to the quantity of data dramatic increase in information system, resulting in system overload and response time extension, and thereby affecting the user experience. How to extract valuable information from the mass of data become a hot topic, thus the research of business intelligence, data warehouse and other technology is increasingly being attention.In order to improve performance and response time in OLAP system, we design a cache mechanism in ROLAP client. The core concept is:This mechanism can selectively save query results and dimension expressions and measure expressions in query statement into the cache; In dealing with the follow-up query, reproduce the query, to avoid get duplicate data from the server, thereby reducing query time; In order to further improve the hit rate of the cache, thereby further improving the overall query performance, using the Incremental Learning Naive Bayes algorithm analysis dimension name of queries involved to decide whether to cache query results. The same idea is also used for cache replacement algorithm. This mechanism uses the in-memory database to simply and easily manage the cached data.In order to verify the validity of this caching mechanism, SSB test model was used in the experiment. From the average query time, buffer cache hit ratio and data volume of the query sent to data warehouse, it was proved that the caching mechanism is effective.