If you missed the presentation you can find the recording and materials here.

Abstract:

This webinar presents IBM Db2's latest advancement in AI-powered
query optimization, introducing AI powered join cardinality estimation,
new in version 12.1.3. Building upon the Single Table Cardinality Model
released in Db2 12.1, this enhancement addresses critical limitations in
traditional join cardinality estimation by leveraging machine learning to
understand complex relationships between local and join predicates
across tables.
Key features and benefits include:
- Reliable performance: some queries execute up to 10 times
faster
- Minimal Resource Consumption: reuse of existing models so that
no new models required
- Automation: reduces query tuning and eliminates the need for
statistical views
In this webinar we will discuss in detail:
- The importance and challenges of cardinality estimation
- How Db2 uses neural networks to improve join cardinality
estimation
- Our internal performance test results on TPC-DS and SAP BW
workloads
- How better cardinality estimates, result in improved access plans
and superior performance without any tuning required.

Speakers:

Lori Strain

Lori Strain is an expert in Db2 Optimization. She has worked in database optimization since
joining IBM in 1989, was instrumental in the development of the original cost model for
Db2 version 2.0 and is very excited to be part of the team working to build the next
generation of data optimization technology based on AI.

Nick Ostan is a software developer and deep learning researcher in the Db2
Query Compiler & AI Optimizer group. Nick’s studied biochemistry and
biophysics at the University of Toronto before transitioning to data science
for antibody engineering, and eventually purely to deep learning at IBM.

Lakshmi Tirunelveli Radhakrishnan

Lakshmi TR is a software developer in the Db2 Query Compiler & AI Optimizer team. She
has been closely involved in the AI Query Optimizer initiative, contributing to the
integration of AI into Db2’s query optimization. Her work focuses on building the
infrastructure that supports both the optimizer and ML components.