International Business Machines Corporation

Servicenow, Inc.

Rocket Software, Inc.

Bryan Frederick Smith

David R Schwartz

James Z Teng

James Alan Ruddy

Fen-Ling Lin

John Marland Garth

Yun Yu Wang

Joel Frank Farber

Holger Karn

Akira Shibamiya

Hong Sang Tie

Tammie Dang

Paramesh Sampatrai Desai

Teresa Lynn Leamon

Donald Allan Weil

Ted Eric Blank

Kathryn Ruth Zeidenstein

Kou Horng Yang

Balakrishna Raghavendra Iyer

Curt Lee Cotner

You-Chin Gene Fuh

James Zu-Chia Teng

Torsten Steinbach

Donald J Haderle

John Joseph Campbell

William Robert Bireley

Randy Mitchell Nakagawa

Peter Hoa

Christian Michel

Jerome Quan Wong

Annie S Tsang

Julie Chen

Maureen Mae McDevitt

Beth Rivenes Moore

Kouhorng Allen Yang

David R Betten

Timm Zimmermann

David Warner

Michael John Bracey

Koshy John

Darren Benjamin Swank

S Venkatesh Gopal

Craig Leonard Sutton

James R Porell

Dino Carlo Tonelli

The invention relates to a method and apparatus for recording and maintaining stored information system object relationship information. Information contained within a stored information system (including system catalogs, referential constraints, triggers, table hierarchies, column references, indexes, stored program packages, system catalogs, stored procedures, stored queries, log/trace files of dynamically executed code, etc.) are searched to identify dependency relationships between objects. This object relationship information is stored and maintained in an information base. Information within the information based may be organized based upon subsets of objects that support a common application, service, or capability. Object relationship data is used to facilitate stored information system administration activities such as generating/restoring from archives, physically/logically segmenting data objects, replicating data objects; optimizing the distribution of data in distributed data system; and estimating operational capacity.

Method and apparatus for recording and managing data object relationship data

A method, apparatus, and computer program product are disclosed for managing query execution plans for a workload. In one embodiment, the method includes gathering an initial set of performance characteristics generated from a database management system (DBMS) executing a database workload according to one or more query execution plans (QEPs). The database workload includes one or more database statements. The method determines a QEP change for one or more QEPs of the database workload. In addition, the method gathers a subsequent set of performance characteristics generated from the DBMS executing the one or more QEPs of the database workload influenced by the QEP change. Furthermore, the method determines a performance difference created by the QEP change based on the initial set of performance characteristics and the subsequent set of performance characteristics and responds to the performance difference in accordance with a monitoring policy.

Automatic query execution plan management and performance stabilization for workloads

Techniques are provided for tuning work. In response to identifying a trigger for a work, automatic tuning is invoked. The trigger indicates that a performance goal for the work has been missed. The work is analyzed. A tuning operation is selected for the work based on a tuning history. The selected tuning operation is implemented for the work.

Tuning of work to meet performance goal

An apparatus, system, and method for determining the maximum supported degree of parallel sort operations in a multi-processor computing environment. An allocation module allocates a minimum number of sort files to a sort operation for each data source that participates in the parallel sort. The allocation module attempts to allocate sort files of one-half the sort operation data source file size, and iteratively reduces the sort file size requests in response to determinations that sort files of the requested size are not available. After allocation, a parallel operation module determines whether there is sufficient virtual storage to execute the sort operations in parallel. If there is not, the parallel operations module collapses the two smallest sort operations, thereby reducing the degree of parallelism by one, and repeats the request. The parallel operation module repeats the process until the sorts are executed or the process fails for lack of virtual storage.

Apparatus, system, and method for deterministic file allocations for parallel operations

A method, computer program product, and system for triggering an update of statistics data used by a query optimizer of a relational database management system for determining an optimal execution plan for executing a received query request on a relational database is provided. A system dependent nominal performance value for executing the query request on a database system is determined. Additionally, an average actual performance value is determined by processing the query request. If a comparison between the average actual performance value and the system dependent nominal performance value yields a result which is greater than a first threshold value, an update cost value and a cost saving value are determined. If a comparison between the cost saving value and the update cost value yields a result which is greater than a second threshold value, the statistics data is updated.

Self-healing RDBMS optimizer

A method, apparatus, and article of manufacture accesses a dataset using an unsupported access method. A first request to access a dataset is intercepted. The first request is associated with a first data structure that specifies a first access method. The first data structure is replaced with a second data structure that specifies a second access method which is different from the first access method. The dataset is accessed in accordance with the second access method of the second data structure. In another particular embodiment, the first access method is the basic direct access method, and the second access method is any of the basic sequential access method and the queued sequential access method. In yet another embodiment, the dataset is an extended format physical sequential dataset.

Accessing a dataset using an unsupported access method

A method and article of manufacture, implementing the method, allocates space for a dataset. The dataset has an initial area and zero or more additional allocated areas to provide space for storing the dataset. The size of a new additional area is determined. The new additional area is associated with a new area number, and the size of the new additional area is based on the new area number. Additional space for the dataset is allocated based on the size of the new additional area.

Computer-implemented method for allocating new additional area for the dataset in storage based on the size of the new additional area wherein if the new area number does not exceed clipping threshold, the size of a new additional area being greater than the size of each previously allocated additional area of the dataset

A method, apparatus, and article of manufacture for a computer-implemented apportioning system. Work is distributed to processors in a multi-processor system. Initially, during bind-time, a scaling factor is determined for each processor. The scaling factor represents relative processing power in relation to each other processor. Then, portions of a total amount of work are distributed to each processor based on the determined scaling factor of that processor and a determined amount of work for an average processor.

Apportioning a work unit to execute in parallel in a heterogeneous environment

A method, apparatus, and article of manufacture for a computer implemented groupby aggregation system. Groupby aggregation is performed when a statement is executed in a database stored on a data storage device, which is connected to a computer. Initially, data elements are input into a tournament tree from an input sequence. Then, a winner of a tournament in the tournament tree is identified. The winner is recycled into the tournament tree.

Efficient groupby aggregation in tournament tree sort

A method, apparatus, and article of manufacture for a computer implemented rebalancing system. Partitioned data is rebalanced in a database stored on a data storage device connected to a computer. Range values are redefined for each partition. Next, the data is reordered into the redefined ranges for the partitions.

Rebalancing partitioned data

A system and method for pacing, or controlling, the processing of multiple producers when a consumer requires results from the producers in natural order. This invention regulates the use of system resources between the producers to ensure that the required results are available to the consumer in natural order with minimal waiting and to prevent unneeded advanced processing by the producers. This invention implements a buffer structure such that each producer writes its results to an associated buffer. Each producer compares its buffer's percentage complete against a next and previous producer's buffer. If a producer produces results too rapidly, the producer suspends itself until it is resumed by the consumer or the previous producer. The consumer reads the results from the buffers in producer order.

Pacing of multiple producers when information is required in natural

A multiprocessing system forms a data structure, such as by loading reorganizing, or recovering, while concurrently collecting various statistics about the data structure. The data structure may comprise tables and/or indices, for example. A first processing unit forms the data structure by assimilating data from one or more data sources into data rows, storing the rows in a buffer, and copying the rows from the buffer to the data structure. Concurrently with the forming step, the same or a second processing unit retrieves the rows from the buffer and applies a predetermined analysis to the rows to formulate statistics regarding the data structure.

System for data structure loading with concurrent statistical analysis

A method, apparatus, and article of manufacture for a computer implemented recover/build index system. The recover/build index system builds a database index for a database file by scanning partitions of the database file in parallel to retrieve key values and their associated record identifier (rid) values. The recover/build index system then sorts the scanned key/rid values for each partition in parallel. Next, the recover/build index system performs one or more merges on the sorted key/rid values from all of the partitions to generate a single key/rid value stream. Finally, the recover/build index system builds the index using the single key/rid value stream.

High performance recover/build index system by unloading database files

A system and method for a computer based system having one or more consuming DBMSs and one or more producing DBMSs such that the producing DBMSs produce workfiles for the consuming DBMSs. A producing DBMS uniquely identifies the bufferpool pages comprising a workfile, uniquely identifies the workfile, and flushes only the bufferpool pages corresponding to the workfile to a storage device. The producing DBMS further notifies the consuming DBMS that the workfile is available and receives a reply from the consuming DBMS indicating that the consuming device retrieved the data from the storage device. The consuming device receives notification from the producing DBMS, retrieves the workfile, and replies to the producing DBMS that it has retrieved the workfile.

Bufferpool coherency for identifying and retrieving versions of workfile

Growing community of inventors

Morgan Hill, CA, United States of America

Bryan Frederick Smith