Electrical computers and digital processing systems: multicomput – Computer network managing – Computer network monitoring
Reexamination Certificate
1998-12-07
2001-08-28
Black, Thomas (Department: 2171)
Electrical computers and digital processing systems: multicomput
Computer network managing
Computer network monitoring
C709S224000, C709S248000, C707S793000, C707S793000
Reexamination Certificate
active
06282570
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates in general to computer-implemented database management systems, and, in particular, to monitoring a large parallel database.
2. Description of Related Art
A database is a collection of stored data that is organized as tables. A table consists of rows and columns of data. The rows are formally called tuples. A database will typically have many tables and each table will typically have multiple tuples and multiple columns. The tables are usually stored on direct access storage devices (DASD) such as magnetic or optical disk drives for semi-permanent storage.
Databases generally have indexes, configuration files, and transaction logs. Database indexes contain a list of keywords for identifying a unique tuple. A database configuration file contains configuration information about the database and information about the database manager. Transaction logs contain a record of all transactions performed on the database.
A parallel database partitions a database over one or more nodes, e.g., from one to hundreds. A database partition is a part of a database. Database partitions have their own data, indexes, configuration files, and transaction logs. In a parallel database, the data in a table can be contained in one or more database partitions. Typically, when a table is on multiple partitions, some of its tuples are stored in one partition and some of its tuples are stored in other partitions. Such a division of a table is called horizontal partitioning.
Database performance is usually important to users. As used herein, the term user is a generic term for a computer user or an enterprise that maintains large amounts of data. Users may monitor the database performance of both non-parallel databases and parallel databases. Monitoring a database typically involves acquiring performance information from an interface exposed by a database instance and evaluating the information for threshold violations or “out of norm” conditions.
Monitoring a non-parallel database generally includes collecting performance statistics of a database. The performance statistics can be used to calculate a performance value for a non-parallel database, using, for example, a predefined formula. The performance value is then compared to a predefined threshold value. When monitoring the performance of a parallel database, some monitors collect performance data from a node and provide the user with a single performance value. Other monitors, referred to as summary monitors, access performance data from multiple nodes. These monitors are typically used when data from multiple nodes is essential to obtaining a meaningful result. For example, data from multiple nodes may be essential when calculating an average value or a maximum value. Hereinafter, the term monitor refers to a summary monitor.
Oftentimes, the monitor must collect the performance data from multiple nodes within a predefined time limit. During the time limit, the monitor may not be capable of collecting statistics from all of the partitioned databases. Overhead will be produced at the monitor to enable the collection of the monitoring statistics.
Turning on the performance statistics also reduces the performance of the database manager and lengthens the time of performing a transaction. Therefore, there is a need for a parallel database monitoring system that returns performance data to a user within a specified time limit and minimizes overhead.
SUMMARY OF THE INVENTION
To overcome the limitations in the prior art described above, and to overcome other limitations that will become apparent upon reading and understanding the present specification, the present invention discloses a method, apparatus, and article of manufacture for monitoring performance of a parallel database in a computer.
In accordance with the present invention, the parallel database is stored on a data storage device in the computer. Groups of database nodes are identified. Collection time periods for collecting performance statistics from the identified group of database nodes are determined. Performance statistics are periodically collected from a subset of each identified group of nodes during collection time periods. The collected performance statistics are cached and re-used when collecting performance statistics from one or more groups of database nodes in a succeeding collection time period.
An objective of an embodiment of the present invention is to increase the precision of monitoring results returned to a user after successive iterations of monitoring. A further objective of an embodiment of the present invention is to return the best results possible within a defined time limit and to minimize overhead across all nodes.
REFERENCES:
patent: 4463424 (1984-07-01), Mattson et al.
patent: 5325530 (1994-06-01), Mohrmann
patent: 5497486 (1996-03-01), Stolfo et al.
patent: 5524253 (1996-06-01), Pham et al.
patent: 5579515 (1996-11-01), Hintz et al.
patent: 5649198 (1997-07-01), Shibata et al.
patent: 5870608 (1999-02-01), Gregory
patent: 5935216 (1999-08-01), Benner et al.
patent: 5938729 (1999-08-01), Cote et al.
patent: 5958009 (1999-09-01), Friedrich et al.
patent: 0 646 883 A1 (1995-05-01), None
patent: 09179834 (1995-12-01), None
patent: WO 96/26488 (1996-08-01), None
IBM Technical Disclosure Bulletin, “Method for Storing Large Objects in a Relational Database”, 35(4A):72-75, (Sep. 1992).
IBM Technical Disclosure Bulletin, “Effective Approach to Query I/O Parallelism Using Sequential Prefetch and Horizontal Data Partitions”, 36(09A):541-543 (Sep. 1993).
IBM Technical Disclosure Bulletin, “Client-Server Algorithm for Distributed Libraries”, 35(2):352-356 (Jul. 1992).
IBM Technical Disclosure Bulletin, “Determining Database Object Authorization”, 32(9B):335-337 (Feb. 1990).
Iyer, B., et al., “Percentile finding algorithm for multiple sorted runs”, Proceedings of the Fifteenth International Conference on Very Large Data Bases, Amsterdam, Netherlands, 1 page (1989).
Frieder, O., et al., “Data distribution and query scheduling policies for a cube-connected multicomputer system”, Second International Conference on Supercomputing Proceedings, San Francisco, CA, 1 page (1987).
Leung Tony Kai-Chi
See Siu Nang
Wong Stan
Wu Cheung-Yuk
Black Thomas
Chen Te
Foerster Ingrid
International Business Machines - Corporation
Komanduri Janaki
LandOfFree
Monitoring a large parallel database through dynamic... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Monitoring a large parallel database through dynamic..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Monitoring a large parallel database through dynamic... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2522610