If the file that you
want to process contains a large amount of data, you might achieve
better performance by dividing that data among several smaller files.
Many of the tools that collect raw data have controls that enable
the user to split the data among several smaller files. For example,
tools that extract SMF data have this capability. For the database-oriented
collectors, you can decide to run the staging job more frequently.
SMF data is typically
collected and written to an output file by using an IBM utility called
IFASMFDP. (This utility is also known as the SMF data set dump program.)
It can be used to produce multiple output SMF files. With only a single
pass of the input, it can select the data according to SMF record
types that are specified in the SMF dump program parameters. For example,
you can generate three separate SMF files: one with DB2 data, one
with CICS data, and one with everything else. Thus, you could execute
three staging jobs concurrently (to read each of the three SMF files)
and then execute their associated aggregation jobs.