|
|
Thesis: "Dynamic Tuning
Algorithms for Multiple Buffer Pools in a DBMS"
|
|
Abstract (full text
not available)
|
|
In order to meet the performance
goals of different applications, database administrators (DBAs) must adjust
a number of low-level performance "knobs". As the complexity
and diversity of data types and database workloads increase, manual tuning
by DBAs becomes almost impossible. There is a need for database management
systems (DBMSs) to perform automatic tuning, based on high-level performance
goals provided by the DBA.
Goal-oriented resource management is the capability of a DBMS to adjust
its low-level configuration parameters in order to achieve predefined high-level
performance goals. Self-tuning algorithms achieve this. They
detect any violations of performance goals and dynamically reallocate database
resources to meet those goals.
A self-tuning algorithm for multiple buffer pools is described in this
thesis. Buffer pools are a key resource in a DBMS. The algorithm
implements the concept of goal-oriented resource management by reallocating
the buffer pool resources to meet the performance goals of various on-line
transaction classes. All the experimental results presented in this
thesis were based on DB2/UDB and the TPC-C benchmark.
|
|