DB2 for IBM i Performance
Lundell, Mattias (2020)
Lundell, Mattias
Åbo Akademi
2020
Julkaisu on tekijänoikeussäännösten alainen. Teosta voi lukea ja tulostaa henkilökohtaista käyttöä varten. Käyttö kaupallisiin tarkoituksiin on kielletty.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi-fe2020062445622
https://urn.fi/URN:NBN:fi-fe2020062445622
Tiivistelmä
The purpose of this master's thesis is to help programmers working with DB2
for IBM i make up-to-date design decisions. Much has changed in the industry in
the past 20 years, and sources of informaton on both the Internet and internal
company policies are outdated and misleading.
Underlying concepts of IBM i and DB2 that affect performance are explained. A
framework for benchmarking database programs on IBM i has been created, using
RPG as the programming language and the built-in DB2 database. Its design is
presented, and parameters relatng to the performance of both RPG and SQL are
explored, to establish how accurate benchmarking tests can be set up.
Benchmarks were created for a few practcal scenarios and the results are
analyzed, and compared to existng literature, when available. The
recommendaton of IBM on index column ordering was shown to be suboptmal
when dealing with very large amounts of data. The performance of batch updates
can be improved by creatng a temporary table in order to avoid changing the
same data that the selecton criteria is based on
for IBM i make up-to-date design decisions. Much has changed in the industry in
the past 20 years, and sources of informaton on both the Internet and internal
company policies are outdated and misleading.
Underlying concepts of IBM i and DB2 that affect performance are explained. A
framework for benchmarking database programs on IBM i has been created, using
RPG as the programming language and the built-in DB2 database. Its design is
presented, and parameters relatng to the performance of both RPG and SQL are
explored, to establish how accurate benchmarking tests can be set up.
Benchmarks were created for a few practcal scenarios and the results are
analyzed, and compared to existng literature, when available. The
recommendaton of IBM on index column ordering was shown to be suboptmal
when dealing with very large amounts of data. The performance of batch updates
can be improved by creatng a temporary table in order to avoid changing the
same data that the selecton criteria is based on