Introduction to Teradata DSA | DSE | DSU - Introduction to Data Stream Architecture - BAR - Data Stream Architecture - Data Stream Extensions

Teradata® DSA - DSE for IBM Spectrum Protect Installation, Configuration, and Upgrade Guide

Data Stream Architecture
Data Stream Extensions
Release Number
November 2022
English (United States)
Last Update
Product Category
Teradata Tools and Utilities

Teradata® Data Stream Architecture (DSA) enables you to back up and restore Teradata system data. DSA is optimized for Teradata MPP Architecture. It integrates with the Teradata® Viewpoint portlets: BAR Setup and BAR Operations. The portlets provide user interfaces to Teradata DSA that are similar to other Teradata ecosystem components. This integration uses Viewpoint account management features and enhances usability. Teradata DSA also provides a command-line utility that you can use to configure, initiate, and monitor backup and restore jobs.

Data Stream Extensions and Data Stream Utility

Beginning with DSA 15.10, the product has been bundled into two components: Data Stream Extensions (DSE) and Data Stream Utility (DSU). Both components offer BAR portlet and command line functionality.
  • DSE adds support for third-party backup applications. DSE offers advanced enterprise backup tools, such as scheduling, retention policies, archiving and allows customers to backup up directly to tape. DSE offers these backup targets:
    • Disk file systems
    • Veritas NetBackup
    • IBM Spectrum Protect
  • DSU does not offer third-party backup application support. DSU is a solution offered for sites without a need for the extended footprint offered by Teradata DSE. DSU is also used for public cloud solutions. DSU offers these backup targets:
    • Disk file systems
    • Dell EMC Data Domain
    • Amazon S3
    • Azure Blob
    • Google Cloud

    In a typical DSU use case, the DSA Network Client (ClientHandler) is installed on the Teradata nodes, the DSC server is provided in a VM format, and a simple NFS environment is set up for use as a storage location for the backup files. A managed storage server can also act as a host server to the NFS environment if needed. When a Data Domain unit is used, EMC Data Domain Boost for DSU (DD Boost) allows a direct connection to the unit without using a third-party backup application.

    Note that DSA consumes a certain amount of memory on the Teradata nodes. The amount of consumption depends on factors such as the following:
    • Number of objects in the job plan
    • Configuration, including how many nodes and many streams
    • Throughput

    For an estimate you can use the following:

    Fixed consumption 2MiB + (Number of streams per node x 4MiB)

    Number of AMPs per node x 15MiB

    Variable consumption With 80K objects, consumes around 1-2GiB per node (not constant consumption):

    20MiB + (Number of objects x 400) + ((Total number of streams x 400) x 3)

    For example, with 80K objects, 4000 total streams ~60MiB

Server Functionality

Server functionality includes the following servers:
  • DSC server, which controls all BAR operations and is a part of all configurations. A DSC server must have the Data Stream Controller (DSC) installed.

    Teradata DSC can be installed on a physical server, AWS, Azure, Google Cloud, or a VM (Teradata DSC on VMware) and back up and restore data from and to the database on-premises, in AWS, Azure, Google Cloud, or VMware.

  • Media server (physical or logical), which writes to the target storage device. A media server must have the DSA Network Client (ClientHandler) installed.

A machine in a DSA configuration can include different types of server functionality. For example, the managed storage server in a DSU configuration functions as disk storage, the DSC server, and a media server. In another configuration, the DSC server could be a standalone server.

Backup Solutions

DSA backup solutions can include any of the following:
  • Dell EMC Data Domain
  • Quantum Tape
  • Disk file system
  • Third-party backup application software, such as NetBackup or IBM Spectrum Protect
  • Amazon S3
  • Azure Blob
  • Google Cloud Platform