Scalable Multimedia Data Warehouse for 21st Century
Felipe Carino
The talk consists of an overview of Teradata's vision about Data Ware-
house, Data Mart, very large databases and parallel systems. Real
cases are used to illustrate points.
- Teradata Data Warehouse/Data Mart Concept: central repository that
stores all raw (scalar) data.
- Growing Pains: experiences growing and developing Data Warehouses.
- Wal*Wart: Case Study on how they use a 7-24 Terabyte Database. I
will explain how Wal*Mart process engineered the Retail Model by
using information as a strategic asset and weapon. I believe this
is the world's largest commercial data warehouse.
- Other Case Studies: I will describe the importance of information
analysis.
- Massively-Parallel Processing Platforms (MPP) and Teradata RDBMS:
analyzes how a Teradata relational database executes queries in
parallel with an MPP system.
The second logical part of the talk describes the changes and new
challenges that object/relational evolution introduces to the Data
Warehouse/Data Mart concepts.
- Object/Relational Database Evolution: changes to Data Warehouse/Data
Mart. I shall describe Teradata Multimedia Database System (MDBS).
- Medical Knowledge Bank: a pioneering 5-year, multi-company NIST-
ATP effort to automate medical records, diagnosis and treatment
end-to-end. This project uses our parallel object/relational
MDBS database.
- Parallel UDF issues: executing UDFs on an MPP platform.