The Reformation Of Database Decision Support Testing System And The Analysis Of InnoDB Index

With the extensive application of technique of computer in every trade,the data has already become precious resources for various organizations.The database management system becomes the important foundation coresoftware technique.Currently the general database product is placed basiclly in the blankcircumstance. The Decision Support Test System will be benefit to ahomemade database production. The system can be applied to design,develop, upgrade and recursively test the database production. The reading ofthe open source database source code contribute to understanding thetechnique of a database and provide the necessary technical support for thehomemade database.This paper is constitute by two parts, the first part mainly introduces thereformation that Database Decision Support Test System; the mainintroduction of the second part is to analyze the directory structure and sourcecode of index processor subsystem of the InnoDB module of open sourcedatabase MySQL5.0. These two parts works are all at for develop thedomestic database to do the technique preparation.The Decision Support Test System is a decision support benchmark. Itconsists of a suite of business oriented ad-hoc queries and concurrent datamodifications. This benchmark illustrates decision support systems thatexamine large volumes of data, execute queries with a high degree ofcomplexity, and give answers to critical business questions.The Decision Support Test System is the subsystem of RedPTS ( theRedBase Performance Testing System). The technique of it comes from theTPC-H and DBT3. The Transaction Processing Performance Council (TPC)Benchmark Specification H (TPC-HTM) is a decision support benchmark. TheOSDL Database Test 3 (OSDL-DBT-3) test kit is a simplified derivative ofthe TPC-H.In actual test process, the DBT3 has many problems.For example:TheDBT3 provides only an easy to use decision support workload of businessoriented ad-hoc queries to stress the Linux operating system for SAP DB andpostgres. But the decision support benchmark support two driver/SUTconfigurations, a "host-based"and a "client/server"configuration。Aim atthese weaknesses, the Database Decision Support Test System that carry outthem through reformating. We implement the test process for redbase andoracle on two driver/SUT configurations. The tested and reported configuration(s)of the Decision Support TestSystem is composed of a driver and a system under test (SUT).The driversubmits queries to SUT. The SUT executes these queries and replies to thedriver. The driver presents the workload to the SUT. The driver is a logicalentity that can be implemented using one or more programs, processes, orsystems and performs the functions. The host system(s) or server(s) includinghardware and software supporting access to the database employed in theperformance test. Ten members in our group spent half a year in analyzing theimplementation of InnoDB storage engine in MySQL 5.0 alpha Open Sourcedatabase. The MySQL software delivers a very fast, multi-threaded, multi-user,and robust SQL (Structured Query Language) database server. MySQLServer works in client/server MySQL, the most popular Open Source SQLdatabase management system, is developed, distributed, and supported byMySQL AB. The MySQL Server design is multi-layered with independentmodules. InnoDB is essentially a stand-alone database management system。InnoDB provides MySQL with a transaction-safe (ACID compliant) storageengine with commit, rollback, and crash recovery capabilities. InnoDB doeslocking on the row level and also provides an Oracle-style consistentnon-locking read in SELECT statements. These features increase multi-userconcurrency and performance. There is no need for lock escalation inInnoDB because row-level locks in InnoDB fit in very little space. InnoDBalso supports FOREIGN KEY constraints. The paper introduces the fuction and principle of InnoDB index inMySQL 5.0 alpha Open Source database. The index and storage managerinterrelate with each other. If you total up the sizes of the C files, you'll see that \btr is thesecond-largest file group in InnoDB. This is understandable becausemaintaining a B-tree is a relatively complex task. Luckily, there has been a lotof work done to describe efficient management of B-tree and B+-treestructures.