Teradata Tuning

1 Module 1 Lesson Intermediate

About this course

Teradata is one of the most sophisticated, high-performance database systems for the data warehouse sector. But as with all shared-nothing, massively parallel processing systems, it is necessary to use the available resources to efficiently achieve optimal performance. In this Teradata course, we share over 25 years of knowledge and experience in SQL tuning.

We describe in a simple and comprehensible way all the information that must be considered when one is called to improve anything from a single badly-performing query up to the entire DWH workload of overnight batch processes failing to meet the service level agreements. The Teradata course follows the full lifecycle of a Teradata DWH, from the architecture and design to optimizing data access paths and execution plans. Anyone involved in the development, administration, or database design will benefit from the tips and tricks presented in this course.

Course Structure

1 Lesson

The Approach of Teradata

In this module, we visit all the building blocks we will need to explore performance tuning in the next chapters. These are the topics of parallel database architectures; the shared-nothing architecture of Teradata; hardware and software components of Teradata, including cabinets, cliques, clusters, nodes, hot-standby nodes, Parallel Database Extensions, Parsing Engine (PE), BYNET, Access Module Processor (AMP), VProc, and Database File System; database objects; storing physical rows in an AMP VDisk, including the Teradata parallel hash algorithm, AMP hash maps, row hashes, physical rows, data blocks, VDisk logical sectors and cylinders, and the Master and Cylinder Index; logical row- and columnar orientation; and DBMS services, including the fallback AMP hash maps, Teradata Intelligent Memory, transactions, permanent and transient journals, WAL protocol, locks and deadlock handling, sessions, and workload management.