Cache Research OLAP Cube-based Suppliers And Implementation

Today known as the age of the "information explosion", a large number of data is accumulated in telecom, banking, online e-commerce sites and other data-intensive enterprises. Through the effective use of the client cache, we can improve the efficiency of user queries. Today, there exist a lot of query cache technologies, such as semantic caching, objects caching and so on. Semantic caching means that the query results and the related query unified stored in the cache. In this paper, it does propose a new kind of cache technology, an OLAP cache model based on quotient cube. The cache model has the following advantages:1) it can store the query results in cache;2) it saves cache space through avoiding redundancy store;3)compared with other technology, it takes less response time and has higher hit ratio.Firstly, this paper proposes an OLAP cache model based on quotient cube, then implements the system based on this model, last but the least, does many experiments with this model. These results of experiments verify the validity and accuracy of this cache model.The main contributions of the paper’s work are as follows.First, based on the different queries submitted by users, it generates the equivalence class of quotient cube in cache dynamically. It will query in cache at first when users submit the queries. If there doesn’t exist equivalence class which contains the matched query item, it will query the database, and get relative aggregation value based on aggregation function. Finally, it produces the equivalence class of quotient cube in cache;Second, producing the equivalence classes of quotient cube with the holistic measure function and not monotone measure function This is the first highlight point in this paper;Third, researching and using the relation of roll up and drill down among the equivalence classes, it dose further improve the cache query efficiency by the prune methods. For there are some roll up and drill down relations among these classes, these classes can composite a algebraic lattice. This is the second highlight point in this paper;Four, researching the relation among cache hit ratio, response time, cache size, cache replacement policies and query submit by users, test and verify the OLAP cache.