公开(公告)号：US10585872B2

公开(公告)日：2020-03-10

申请号：US15836987

申请日：2017-12-11

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION

Inventor： Thomas F. Boehme ,  Andreas Brodt ,  James L. Finnie ,  Knut Stolze

IPC: G06F16/22

Abstract: Reorganizing a data table to improve analytical database performance is presented. Attribute value information for data blocks of the data table is determined. One or more queries having respective query predicates involving the attribute may be received and processed. Data blocks having attribute value information matching a query predicate are selected. The selected data blocks are evaluated for the query predicate of the query. For each part of a set of the parts a false positive rate may be determined. The false positive rate may indicate the fraction of data blocks of the selected data blocks of the part, that each contains less than a determined maximum number of rows that satisfy the query predicate. In response to determining for each part of the set of the parts that the determined false positive rates of the part meet a predefined condition sorting the part by the attribute.

公开(公告)号：US10282349B2

公开(公告)日：2019-05-07

申请号：US14836042

申请日：2015-08-26

Applicant: International Business Machines Corporation

Inventor： Gordon A. Booman ,  James L. Finnie ,  Oliver Koeth ,  Sunil K. Sarin

IPC: G06F17/30 ,  G06F16/21 ,  G06F16/22 ,  G06F16/23

Abstract: In an approach to storing data elements in a database, a database management unit provides a table-mapping file for enabling access to a table of a plurality of tables in a storage medium, the table-mapping file comprising a plurality of map records of a single table, a map record comprising information indicative of the table, a column of the table, a data section of the table, and a range of values of data elements included in the data section of the column. The unit receives a data element at the database via a communication interface, determines the table, column, and data section the data element is to be stored in, and stores the data element in the section file corresponding to the determined table, column and data section. The unit then updates the table-mapping file corresponding to the table, columns and data section the data element has been stored in.

公开(公告)号：US20180101555A1

公开(公告)日：2018-04-12

申请号：US15836987

申请日：2017-12-11

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION

Inventor： Thomas F. Boehme ,  Andreas Brodt ,  James L. Finnie ,  Knut Stolze

IPC: G06F17/30

CPC classification number: G06F16/2282

Abstract: Reorganizing a data table to improve analytical database performance is presented. Attribute value information for data blocks of the data table is determined. One or more queries having respective query predicates involving the attribute may be received and processed. Data blocks having attribute value information matching a query predicate are selected. The selected data blocks are evaluated for the query predicate of the query. For each part of a set of the parts a false positive rate may be determined. The false positive rate may indicate the fraction of data blocks of the selected data blocks of the part, that each contains less than a determined maximum number of rows that satisfy the query predicate. In response to determining for each part of the set of the parts that the determined false positive rates of the part meet a predefined condition sorting the part by the attribute.

公开(公告)号：US20170322962A1

公开(公告)日：2017-11-09

申请号：US15146176

申请日：2016-05-04

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION

Inventor： Thomas F. Boehme ,  Andreas Brodt ,  James L. Finnie ,  Knut Stolze

IPC: G06F17/30

CPC classification number: G06F17/30339

Abstract: Reorganizing a data table to improve analytical database performance is presented. Attribute value information for data blocks of the data table is determined. One or more queries having respective query predicates involving the attribute may be received and processed. Data blocks having attribute value information matching a query predicate are selected. The selected data blocks are evaluated for the query predicate of the query. For each part of a set of the parts a false positive rate may be determined. The false positive rate may indicate the fraction of data blocks of the selected data blocks of the part, that each contains less than a determined maximum number of rows that satisfy the query predicate. In response to determining for each part of the set of the parts that the determined false positive rates of the part meet a predefined condition sorting the part by the attribute.

公开(公告)号：US20140032765A1

公开(公告)日：2014-01-30

申请号：US13950414

申请日：2013-07-25

Applicant: International Business Machines Corporation

Inventor： Gopi K. Attaluri ,  James L. Finnie ,  Stewart L. Palmer ,  Piotr M. Plachta ,  Garret F. Swart ,  Xun Xue ,  Roger L.Q. Zheng

IPC: G06F9/50

CPC classification number: G06F17/30362 ,  G06F9/5005 ,  G06F9/5038 ,  G06F9/524 ,  G06F9/526

Abstract: In a shared data system comprising one or more primary nodes and a plurality of secondary nodes, global lock manager on a primary node manages locks for shared resources by exchanging an abstract lock state with local lock managers on the secondary nodes. The abstract lock state includes a particular representation of all of the applications on the nodes that are requesting or are granted locks. The exchange of these particular lock states instead of individual requests improves performance by increasing concurrency and reducing off-machine communication. A global deadlock detector on a node detects and resolves global deadlocks, in conjunction with local deadlock detectors on the secondary nodes.

Abstract translation: 在包括一个或多个主节点和多个次节点的共享数据系统中，主节点上的全局锁管理器通过与次节点上的本地锁管理器交换抽象锁状态来管理用于共享资源的锁。 抽象锁定状态包括节点上正在请求或被授予锁定的所有应用程序的特定表示。 这些特定锁定状态的交换而不是单独的请求通过增加并发性和减少机外通信来提高性能。 节点上的全局死锁检测器与辅助节点上的本地死锁检测器一起检测并解决全局死锁。

公开(公告)号：US10922288B2

公开(公告)日：2021-02-16

申请号：US16358758

申请日：2019-03-20

Applicant: International Business Machines Corporation

Inventor： Gordon A. Booman ,  James L. Finnie ,  Oliver Koeth ,  Sunil K. Sarin

IPC: G06F16/21 ,  G06F16/22 ,  G06F16/23

Abstract: In an approach to storing data elements in a database, a database management unit provides a table-mapping file for enabling access to a table of a plurality of tables in a storage medium, the table-mapping file comprising a plurality of map records of a single table, a map record comprising information indicative of the table, a column of the table, a data section of the table, and a range of values of data elements included in the data section of the column. The unit receives a data element at the database via a communication interface, determines the table, column, and data section the data element is to be stored in, and stores the data element in the section file corresponding to the determined table, column and data section. The unit then updates the table-mapping file corresponding to the table, columns and data section the data element has been stored in.

公开(公告)号：US10331670B2

公开(公告)日：2019-06-25

申请号：US15711216

申请日：2017-09-21

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION

Inventor： Thomas F. Boehme ,  Andreas Brodt ,  James L. Finnie ,  Oliver Schiller

IPC: G06F16/22 ,  G06F16/28 ,  G06F16/2455

Abstract: The method may include providing, in accordance with a column-oriented storage technique, the data table as columns corresponding to the plurality of attributes, whereby each column includes a plurality of separate data blocks. The method may also include determining the plurality of records of the provided data table for which a plurality of attribute values of at least one selected column is contained in a plurality of predetermined data blocks. The method may further include determining, for each column of at least a part of the plurality of columns within the determined plurality of records, a plurality of attribute value information descriptive of an associated attribute within the column and providing an indication of the one or more data blocks for which the plurality of attribute value information is determined. The method may also include storing the determined plurality of attribute value information for enabling query processing.

公开(公告)号：US20170323003A1

公开(公告)日：2017-11-09

申请号：US15147983

申请日：2016-05-06

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION

Inventor： Thomas F. Boehme ,  Andreas Brodt ,  James L. Finnie ,  Oliver Schiller

IPC: G06F17/30

CPC classification number: G06F16/2455 ,  G06F16/221 ,  G06F16/289

Abstract: The method may include providing, in accordance with a column-oriented storage technique, the data table as columns corresponding to the plurality of attributes, whereby each column includes a plurality of separate data blocks. The method may also include determining the plurality of records of the provided data table for which a plurality of attribute values of at least one selected column is contained in a plurality of predetermined data blocks. The method may further include determining, for each column of at least a part of the plurality of columns within the determined plurality of records, a plurality of attribute value information descriptive of an associated attribute within the column and providing an indication of the one or more data blocks for which the plurality of attribute value information is determined. The method may also include storing the determined plurality of attribute value information for enabling query processing.

公开(公告)号：US20150019739A1

公开(公告)日：2015-01-15

申请号：US14338452

申请日：2014-07-23

Applicant: International Business Machines Corporation

Inventor： Gopi K. Attaluri ,  James L. Finnie ,  Stewart L. Palmer ,  Piotr M. Plachta ,  Garret F. Swart ,  Xun Xue ,  Roger L.Q. Zheng

IPC: G06F17/30 ,  G06F9/50

CPC classification number: G06F17/30362 ,  G06F9/5005 ,  G06F9/5038 ,  G06F9/524 ,  G06F9/526

Abstract: In a shared data system comprising one or more primary nodes and a plurality of secondary nodes, global lock manager on a primary node manages locks for shared resources by exchanging an abstract lock state with local lock managers on the secondary nodes. The abstract lock state includes a particular representation of all of the applications on the nodes that are requesting or are granted locks. The exchange of these particular lock states instead of individual requests improves performance by increasing concurrency and reducing off-machine communication. A global deadlock detector on a node detects and resolves global deadlocks, in conjunction with local deadlock detectors on the secondary nodes.

Abstract translation: 在包括一个或多个主节点和多个辅助节点的共享数据系统中，主节点上的全局锁管理器通过与次节点上的本地锁管理器交换抽象锁状态来管理共享资源的锁。 抽象锁定状态包括节点上正在请求或被授予锁定的所有应用程序的特定表示。 这些特定锁定状态的交换而不是单独的请求通过增加并发性和减少机外通信来提高性能。 节点上的全局死锁检测器与辅助节点上的本地死锁检测器一起检测并解决全局死锁。

公开(公告)号：US08868755B2

公开(公告)日：2014-10-21

申请号：US13950414

申请日：2013-07-25

Applicant: International Business Machines Corporation

Inventor： Gopi K. Attaluri ,  James L. Finnie ,  Stewart L. Palmer ,  Piotr M. Plachta ,  Garret F. Swart ,  Xun Xue ,  Roger L.Q. Zheng

IPC: G06F15/173 ,  G06F17/30 ,  G06F9/52 ,  G06F9/50

CPC classification number: G06F17/30362 ,  G06F9/5005 ,  G06F9/5038 ,  G06F9/524 ,  G06F9/526

Abstract: In a shared data system comprising one or more primary nodes and a plurality of secondary nodes, global lock manager on a primary node manages locks for shared resources by exchanging an abstract lock state with local lock managers on the secondary nodes. The abstract lock state includes a particular representation of all of the applications on the nodes that are requesting or are granted locks. The exchange of these particular lock states instead of individual requests improves performance by increasing concurrency and reducing off-machine communication. A global deadlock detector on a node detects and resolves global deadlocks, in conjunction with local deadlock detectors on the secondary nodes.