发明名称 Offloaded, incremental database statistics collection and optimization
摘要 A method and associated system for offloading the collection and optimization of database statistics to an external computer system, without using significant resources required by the database system during its normal operation. A processor of a computer system captures and stores information about the database's recent transactions, where the captured information describes how each transaction altered the logical description of the database or revised data stored in the database. The processor uses this captured information to update stored metadata that identifies the database's logical structure, and to further update stored database statistics that describe characteristics of the data stored in the database. These updated statistics are then returned to the production database's optimizer module, which uses them to determine how to efficiently respond to database-access requests.
申请公布号 US9361338(B2) 申请公布日期 2016.06.07
申请号 US201314089855 申请日期 2013.11.26
申请人 International Business Machines Corporation 发明人 Pathak Ramesh Chandra;Rao Suryanarayana K.
分类号 G06F7/00;G06F17/00;G06F17/30 主分类号 G06F7/00
代理机构 Schmeiser, Olsen & Watts, LLP 代理人 Schmeiser, Olsen & Watts, LLP ;Pivnichny John
主权项 1. A method for offloading the collection and optimization of database statistics, the method comprising: a processor of a computer system capturing transaction information of a production database, wherein the production database is stored on a database hardware storage medium, wherein the captured transaction information describes a characteristic of a transaction associated with a stored data element of the production database and with a logical element associated with the production database, wherein an optimizer module associated with the production database comprises software that increases an efficiency of a data-access operation, and wherein the data-access operation is associated with the stored data element and with the logical element, wherein the captured transaction information comprises reading stored transaction data from the database hardware storage medium, where the stored transaction data was stored on the database hardware storage medium, and wherein the stored transaction data is selected from the group comprising: one or more transaction logs of the production database; data generated by the database software module to facilitate an undo function of the production database; and data generated by the database software module to facilitate a redo function of the production database; the processor storing the captured information in a transaction-log repository, wherein the transaction-log repository is stored on a first computer-readable hardware storage medium that is distinct from the database hardware storage medium; the processor updating cumulative statistical data of the production database as a function of the captured transaction information, wherein the cumulative statistical data may be used by the optimizer module to increase the efficiency of the data-access operation wherein the updating further comprises:the processor inferring metadata information from the captured transaction information, wherein the metadata information identifies a characteristic of the logical element; the processor updating a metadata repository as a function of the inferred metadata information, wherein the metadata repository is stored on a second computer-readable hardware storage medium that is distinct from the database hardware storage medium, and wherein the metadata repository stores information that identifies characteristics of a plurality of logical elements associated with the production database; the processor making the cumulative statistical data available to the optimizer module, wherein the making the cumulative statistical data available comprises an operation selected from the group comprising: storing the cumulative statistical data in a third computer-readable hardware storage medium that is distinct from the database hardware storage medium and that may be accessed by the optimizer module; storing the cumulative statistical data in an information repository that is associated with the optimizer module; and communicating the cumulative statistical data to the optimizer module by means of a communications link between the processor and the optimizer module.
地址 Armonk NY US