Oracle: data files include: control files, data files and redo log files, the parameter file, archive, password file, this is the line according to document function division, and all files are binary encoded file, the database algorithm efficiency is greatly improved, because of the uniformity of Oracle file management to parsing and optimization in the process of SQL execution, specify the unified standard:
RBO (rule-based optimizer), the CBO (cost-based optimizer)
Through the choice of the optimizer, and HINT rules, giving the SQL optimization great freedom, CPU, memory, IO resources for all aspects of optimization,
MySQL: one of the biggest characteristic, is free to choose the storage engine, each table is a file, can choose the appropriate storage engine, common InnoDB engine, MyISAM, NDBCluster etc., but owing to the open plug-in storage engines, such as requiring loose coupling relationship between database and engine, which can lead to the consistency of the documents is greatly reduced, in terms of SQL execution optimization, also has some inevitable bottleneck, in the table, query optimization, statistical function and so on is soft rib, and supports only extremely simple HINT,
SQL Server: basic data architecture is the longitudinal division, is divided into: Protocol Layer (Layer Protocol), Relational Engine (Engine), Storage Engine (Storage Engine), and the SQLOS, SQL execution process is step by step a parsing process, the Relational Engine in the optimizer, is based on the cost (CBO), Oracle and its working process are very similar, on cost and support a rich HINT, including: connection, query HINT, HINT,

CodePudding user response:

Huge amounts of data with oracel, cross-platform used mysql, using SQL server on Windows