Research On OLAP Semantic Cache Model

In recent years,the connection between big data and traditional industries has become increasingly close,which has made the traditional industry continue to ignite new vitality,but it also brings some challenges in the analysis and application of data.In big data analytics,OLAP(online analytical processing)and aggregated queries are often used,and these queries have stronger semantics and contextual relationships(such as drill down and up summary)than general SQL queries.However,traditional query caching does not take full advantage of the features of OLAP or aggregated queries.Because in a limited memory space,they neither explore the semantics of OLAP nor the compression of data,especially in the context of big data.This paper proposes an OLAP semantic cache model that uses the semantic relationship between data units in the data cube to facilitate query and compression in the cache.The quotient cube equivalent class is used as the object of cache management in the OLAP semantic cache model.As a caching model,the main work of this paper is as follows: 1)The cache item is dynamically generated in the query,and the method of updating the cache item is found according to the principle of connection equivalence;2)using the data unit known inside the cache item to derive Unknown data unit;3)According to the layer number relationship between the cache item and the query item,a reasonable index is designed to improve the efficiency of query matching;4)The covering capacity of the cache item is calculated by using the principle of inclusion-exclusion as the basis for cache replacement.Compared with traditional query caching,the innovation of this model is that it can compress data while preserving the semantics of data units.In this paper,the real data set and the synthetic data set are used respectively,and the performance of the cache is verified by changing the number of dimensions,the number of tuples,and the inclination of the query items.Compared with the traditional algorithm based on LRU and LFU replacement strategy,in the non-skewed query,the hit rate is increased by about 30% due to the cached semantics,and the hit rate is increased by nearly 40%.In the cache compression experiment,the cache compression rate is about 45%.