Based On Data Warehouse Physical Structure Optimization Research

With the development of science and technology and the actual production needs, data warehouse technology used more widely. Data warehouse system performance is taken seriously more and more, especially for some large enterprise . In the existing data warehouse optimization technique, most focused on a particular data warehouse system related to the SQL statement optimization, and the optimization research of the physical structure for data warehouse is relatively small.This article from the data warehouse physical structure design,combining with demand of practical application,study data warehouse adjustment optimization technique, main work is as follows:1.According to the structure of Shared Pool,the main characteristics of basic framework and existing Bucket partition model, carries on the deep discussion,through the analysis of the divided theory and planning for model, this paper brings forward an improved Shared Pool Bucket partition algorithm,the measures are at the time of division of Bucket,adjust the number of Bucket,to determine the different limits of each part;then the step length divided Bucket was improved, and change Size, will do to reduce the number of small probability Bucket, improve storage relatively small memory block (the Bucket Chunk) quantities.2.Put forward a new optimization algorithm of Buffer cache.According to the management of the latch,significant contention,frequently concurrent access to damage data in memory.The article made each latch to mapping many buckets and increase the number of Bucket,reach use less Latch to manage more bucket,optimization algorithm performance.3.For the PGA, the main research feedback loop algorithm to realize automatic management,and how to optimize PGA's size.In addition,wait for events and data blocks combined to optimize the performance of the data warehouse.The article based on Better Life Commercial Chain Share Co.,Ltd's data warehouse system as an application example,in the optimization application of data warehouse,for the performance of the system,the system data collection, statistics, check,compared and analyzed in detail the relevant data information, which validated the improved algorithm and measures the performance of the data warehouse played optimization function.