Banca de QUALIFICAÇÃO: Eduardo Pingarilho Mendizabal

Uma banca de QUALIFICAÇÃO de MESTRADO foi cadastrada pelo programa.
STUDENT : Eduardo Pingarilho Mendizabal
DATE: 14/06/2024
TIME: 08:00
LOCAL: Remoto via Teams
TITLE:

Using Machine Learning to Adjust Bufferpools in a Relational Database in Mainframe Infrastructure

KEY WORDS:

DBMS, Database, Bufferpool, Machine Learning, Knob Configuration

PAGES: 69
BIG AREA: Ciências Exatas e da Terra
AREA: Ciência da Computação
SUBÁREA: Sistemas de Computação
SPECIALTY: Arquitetura de Sistemas de Computação
SUMMARY:

The buffer pool is an essential component of a Database Management System that stores frequently accessed data pages in memory, improving system performance. It allows for faster data access compared to retrieving data from the disk. As the main memory resource of a database, optimizing the buffer pool can significantly enhance application performance. Properly adjusting its size is crucial for this improvement. The database administrator's responsibility is to configure and manage the buffer pool. However, manual optimization becomes unfeasible in large-scale environments with dynamic and heterogeneous applications, especially when there are multiple buffer pools and instances. This work proposes a resource optimization approach using data mining techniques in response to this need. For this purpose, configuration, and usage data from a corporate database of a large financial sector company were used. Preliminary results are promising, indicating that the techniques employed for optimization are adequate, although further analysis is needed to define the better knob configuration properly.

COMMITTEE MEMBERS:
Presidente - 1706731 - ALETEIA PATRICIA FAVACHO DE ARAUJO VON PAUMGARTTEN
Interno - 2518570 - MARCOS FAGUNDES CAETANO
Interna - 1713597 - MARISTELA TERTO DE HOLANDA
Externo à Instituição - ALAN DEMÉTRIUS BARIA VALEJO - UFSCAR