Databases Reference
In-Depth Information
The product, VM Data Compression, performs compression and expan-
sion dynamically as files are being created and updated or read by applica-
tion programs. Because data compression is automatic and transparent, no
modifications are required to existing application programs. Data compres-
sion is controlled through the use of compression rules stored in a rules
database managed by a rules server virtual machine.
Five compression techniques are offered, including a dynamic data ana-
lyzer that selects the compression technique that yields the best compres-
sion performance with the least consumption of system resources. A trial
utility and compression simulator allows system administrators to evalu-
ate the benefits of data compression without actually writing compressed
data in CMS files. Compression activity is reported in compression statis-
tics records, which are stored in a statistics database managed by a statis-
tics server virtual machine.
For IMS/ESA users, IBM offers IMS Compression-Extended to help reduce
DASD storage requirements for databases, reduce database I/O, and
improve performance. The most efficient compression is accomplished by
exploiting select ES/9000 hardware. However, if there is no compression
hardware installed, MVS compression services simulate the hardware
instructions with software.
To use the hardware data compression capabilities, MVS requires that
compression dictionaries be provided in a specific format. IMS Compres-
sion-Extended provides both a generic dictionary and the capabilities to
create enterprise and custom dictionaries. The generic dictionary immedi-
ately provides the means for good compression, with minimal effort. The
enterprise dictionary maximizes compression for similar enterprise data-
bases. The custom dictionary maximizes compression for a specific data-
base. IMS Compression-Extended links these dictionaries with MVS ser-
vices to create IMS Segment Edit/Compression exit routines.
IMS Compression-Extended is designed around the activities needed to
implement compression:
The Analyze function reads sample data from a specific database, sim-
ulates a dictionary, and reports potential compression statistics.
The Build function uses the same input sources as the Analyze func-
tion to build, verify, and archive a compression dictionary. By linking
this dictionary to the Base Exit Routine, it creates the Edit/Compres-
sion exit routine necessary for compression.
The Compress function provides JCL (job control language) to assist
with the implementation of compression. This function uses the Ed-
it/Compression routine defined in the DBDGEN to reload and com-
press a database after building the Reload JCL.
Search WWH ::
Custom Search