Table of contents
under the product nameDB2 is driven out by IBM for different operating system platforms of the source code and from the characteristics different software for data base management systems. Thus written versions for IBM Mainframes in PL/X ( today's Maschinenbezeichnung exist: zSeries) for the Mainframe operating system z/OS (before times MVS) or VSE, furthermore into the operating system OS/400 integrated a version for IBM avoiding rank systems (today's Maschinenbezeichnung of iSeries), as well as in C written versions for unix -, linux - and windowsbasierte systems.
The versions are current:
- DB2 UDB for z/OS, version 8
- DB2UDB for Linux, UNIX and Windows, V8.2.3
- DB2 UDB for iSeries, version the 5 releases 3
- DB2 server for VSE & VM, V7.4
DB2 administers data in Tablespaces. The system permits at present (2006) maximally 32768 Tablespaces with a maximum sizeof 512 gigabyte per Tablespace. The system can store and administer thus theoretically 16384 Terabyte a comprehensive data set. It supports also binary data types (text, tones, pictures , videos ) apart from the SQL data types. Together with Oracle DATA cousin and ms SQL serverbelongs DB2 to the commercial systems with the largest market shares.
Since February 2006 there is a free version (express C) for Windows and Linux with the following restrictions in relation to the commercial versions:
- max. 2 CCUs (only after bases are counted,several cores and cycles per second thereby) max
- 4 GB main storage this
version is not considered has compared to the commercial variants no restrictions regarding size of the data base and number of users.
DB2 for Unix and Windows becomes with SQL - instructions inthe command line administers or graphically over the control center (db2cc). The administration on Mainframes usually takes place by means of batch jobs, whereby between DB2 utilities (RUN ACT, CoPy, REORG etc.) and DBA is differentiated - jobs (SQL becomes by means of DSNTIAD in a TSO- Backgroundjob accomplished). Smaller work is accomplished often also at the TSO terminal by means of SPUFI under ISPF. In order to obtain during the execution of integrated railway accesses (static SQL) an optimal performance, a so-called becomes. Optimizer assigned, an expert system, which with thatProgram preparation the access to the tables concerned specifies. This is based on the so-called. Statistics, by means of o. A. Tool to be run act periodically updated.
The DB2-EEE in such a way specified (speak “tripolarize I”) is for larger Unix or Windowsumgebungen, whereby the data base partitions overseveral computers (Nodes) to be distributed can. Larger Mainframeumgebungen uses DB2 DATA sharing, whereby the functionality parallel of the Sysplex of the zSeries computers is fully used.
RUN ACT are utility for the production of statistic data over contents of DB2-Tabellen. For example the minimum and maximum values of a column and the Kardinalität of the columns and the table are determined.
These data in the DB2-Systemkatalog, a collection of relational tables, in those are stored DB2 information about all objects, like tables, indices, columns etc., put down (self descriptive, thus „describing “).
The data base management system needs these statistic data, over for thosefrom the user accesses to the DB2-Tabellen realized as optimal an access path as possible to use. Z. B. index accesses are usually senseless, if the entire table contains only few lines; simple reading of all data is then faster. Beside the datafor the access path (Optimizer) in addition, data for the administration are determined, z. B. the quotient of the used memory sides or the compression degrees.
RUN ACT should be implemented always then, if contents of tables substantially changed, orif new indices were put on. Afterwards also the referring DB2-Packages must again be bound with static SQL, since into it the means of access are put down.
The Tool knows for table ranges (Tablespace) or indices to be implemented and knows also in the framework other administrativelyUtilities embedded run (REORG, LOAD).
//RSTAT EXEC DSNUPROC, UID='JCA.RUNSTA', TIME=1440, // UTPROC=, // SYSTEM='V71A', DB2LEV=DB2A //UTPRINT DD SYSOUT=* //SYSIN Dd * RUN ACT TABLESPACE DSNXXX.DSNABCDE TABLE (UNIVERSE) SAM-POLARIZES 25 INDEX (UNIVERSE) SHRLEVEL CHANGES
become DSNXXX.DSNABCDE in the Tablespace herein all tables and indices 25% of the lines examined (SAM-POLARIZE 25). A parallel update by other processes is permitted (SHRLEVEL CHANGES).
REORG is utility for the reorganization of DB2-Tabellen (Unix and Windows version)or Tablespaces (Mainframe). The data are stored in an optimal way on the permanent memory. The optimal way is determined over the cluster index. The memory sides are optimally filled up and lines, which do not stand on their original side(Far or PAGE, Near OF PAGE), the optimum position are shifted.
Utility can work off-lines or on-line. With the off-line variant the data the Clustering index are sorted accordingly and stored into temporary files. The Tablespace is then again put on and the data- now sorted - again into the table range are registered. Subsequently, all indices are again developed.
With on-line variant a new Tablespace („shade copy “) will become put on and the data gradually into the new range sorted to transfer. Intermediate changes become afterwardsfrom the Recovery log trained, whereby a work table (Mappingtable) serves the allocation from old to new internal typographic keys (record Identifiers). If all changes are considered, „a SWITCH takes place “, after which the DB2 accesses from now on the new table range. The old rangeone rejects.
Additionally to reorganizing backup copies (CoPy) and statistic data (RUN ACT) can be determined.
//REORG EXEC DSNUPROC, UID='JCA.REORG', TIME=1440, // UTPROC=, // SYSTEM='V71A', DB2LEV=DB2A //UTPRINT DD SYSOUT=* //SYSIN Dd * REORG TABLESPACE DSNXXX.DSNABCDE
becomes herethe Tablespace DSNXXX.DSNABCDE reorganizes.
the functionality of DB2 can with so-called. Extendern to be extended.
for geographical problems (tide zone analysis etc.)
Object relationally Extender
extension for working with objects upRDBMS basis
in the course of the development of the relational user interface” Structured Query LANGUAGE “(SQL) became IBM-internally the first prototype, which develops system so mentioned R (1975 - 1979). From these experiences became SQL/DS for DOS /VSE develops. Parallel for this the independent line of the Mainframe MVS (today z/OS) was advanced. DB2 was the second available relational data base at the market.
Web on the left of
- official starting side ofIBM to DB2
- the DB2 Cookbook of Graeme Birchall
- German DB2-Client installation guidance for (Gentoo) Linux
- DB2 universal DATA cousin version 8.2 for Linux HOWTO
- graphic data bank pixafe based on DB2 as single place system, Client/servers system and Web version available