Thursday, September 15, 2005
12:00 pm - 1:00 pm
Wean Hall 8220
Continuous Resource Monitoring for Self-predicting DBMS
Administration tasks increasingly dominate the total cost of ownership
of database management systems. A key task, and a very difficult one
for an administrator, is to justify upgrades of CPU, memory and
storage resources with quantitative predictions of the expected
improvement in workload performance. Current database systems are not
designed with such prediction in mind and hence offer only limited
help to the administrator. This paper proposes changes to database
system design that enable a Resource Advisor to answer what-if questions about resource upgrades. A prototype Resource Advisor built to work with a commercial DBMS shows the efficacy of our approach in predicting the effect of upgrading a key resource---buffer pool size---on OLTP workloads in a highly concurrent system.
Eno Thereska is a fourth-year PhD student working with Prof. Greg Ganger. For his PhD he is focusing on how to design self-managing systems. Currently he is working on enabling a large-scale storage system to answer what-if questions about the effect of resource upgrades, service migration and data distribution. He is interested in applying methods from queuing analysis and machine learning to this problem. In the context of self/managing systems, he has also worked on Attribute-Based Learning Environments, which allow a storage system to make educated guesses on how clients will use their files at file creation time and Continuous Reorganization, which attempts to continuously refine the data layout on disks based on user access observations.
Seminar Info Contact:
or visit http://www.pdl.cmu.edu/SDI/