Font Size: a A A

Research On The Index And Recovery Technology Of Java MMDB

Posted on:2020-09-02Degree:MasterType:Thesis
Country:ChinaCandidate:Y WangFull Text:PDF
GTID:2428330575962067Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
With the development of Internet industry,real-time and high-performance transaction processing has gradually become the pursuit direction of applications and management systems.As a database playing an important role in it,whether it has high performance and real-time characteristics is particularly important.Compared with traditional disk resident database,the advantages of memory database are gradually revealed.Java is the most widely used programming language at present.Java project occupies a large share of the market,so it is necessary to study Java MMDB.Through reading a lot of literature and analyzing the current research situation at home and abroad,the author finds that there are some room for optimization and improvement of the key technologies of Java MMDB,which affect the overall performance of Java MMDB.Through the analysis and research of these key technologies,this thesis chooses index technology and recovery technology of Java MMDB to improve and optimize.Firstly,an improved CSB+ tree index structure is proposed in this thesis.To solve the problems of CSB+tree,the intra-node partition is used to expand the capacity of nodes,and the main operation algorithm of the index structure is given.Secondly,a log-driven checkpoint is proposed in this thesis.The improved algorithm is used to reduce the number of logs and make up for the deficiency of the algorithm.Therefore,this thesis proposes an improved log-driven checkpoint algorithm to reduce the number of logs and makes up for the deficiency of log-driven checkpoint algorithm that frequently accesses disks and reduces database performance.Finally,the optimized and improved Java memory database is applied to the actual Java project,which cooperates with disk database.The optimized performance analysis shows that the improved CSB+ tree of Java MMDB index structure proposed in this thesis not only keeps the cache sensitivity of CSB+ tree,but also reduces the number of TLB mismatches.Compared with the original index structure,the improved CSB+ tree has better query performance.Compared with the original algorithm,the improved log-driven checkpoint algorithm proposed for recovery technology reduces the number of logs generated,reduces a large number of invalid disk data updates,and improves the overall time efficiency of Java MMDB.For functions with high real-time performanceand transaction throughput,the Java MMDB can reduce the time consumed in data operation by at least 30%.
Keywords/Search Tags:Java MMDB, Key Technology, Index Technique, Recovery Technology
PDF Full Text Request
Related items