This is the pictorial view of the Teradata Analytical Ecosystem. It provides the foundation to host and analyze all the businesses data. Most customers will not have everything portrayed in this graphic, but may have a subsets as appropriate for their business. Data starts from the source or transactional systems such as sales, inventory, and customer information. The data is then loaded into the primary production integrated data warehouse (EDW) using commercial Extraction and Transformation Tools (ETL) along with Teradata bundled utilities. Backup, Archival, and Retrieval (BAR) is used for backup/archival purposes as well as for restoring systems in case of a system loss. Customers may chose to create Data Labs within the production data warehouse for special purpose applications, testing, or proof of concepts. Customers may also chose to stay in this type of environment for a period of time without the need to add any additional appliances. But overtime customers will add other appliance class systems to complement their integrated data warehouse environment or they may automatically inherit other data warehouses due to consolidations or acquisitions: Test or development appliance (Data Mart Appliance) Special purpose applications on a dedicated dependent data mart (Data Warehouse Appliance) Teradata appliance to support large data volumes (Extreme Data Appliance) Entry level data warehouse (Data Warehouse Appliance) High performance pure solid-state devices (SSD) (Extreme Performance Appliance) Data will be synchronized between the Data Warehouse and the Appliances in a variety of ways and we will discuss the synchronization options later. As customers employ these special purpose systems for their distinct workloads, we call these architectures “dual systems for workload optimization”, which in essence means that the right workload is put on the best platform to achieve the price/performance/value required for that application. Finally, as the business requirements increase demand for high availability, customers will not only increase single system availability, but will choose to employ a dual systems strategy for high availability and disaster recovery. We call these architectures “dual systems for availability”. The most common availability architecture is Dual Active, where two systems are configured in an Active/Active configuration. Another common availability architecture is Active/Standby, followed by an environment which only delivers disaster recovery where the primary production system is coupled with a disaster recovery system at a remote location that is only used in a true disaster. NOTE – although BAR is only shown on two systems in this diagram, BAR should be on all production systems and may be a dedicated or shared infrastructure. BAR was left off the other systems just to keep the slide from getting too busy. Business users access the analytical applications such as CRM and SCM as well as BI solutions which complete the ecosystem by providing subject-based analytics with familiar user interfaces such as dashboards and scorecards and also direct access to the analytical data Finally, the database administrators (DBA’s) and system administrators get a single operational view to the entire Analytical Ecosystem for systems management, alerts, and resource optimization. The resulting Teradata Analytical Ecosystem is a very flexible environment which enables enterprise analytics.
Questions & Answers Now, we welcome your questions. Please state your name and medium or blog, then ask your question. Please use the microphones – just signal Dan and Mike and they’ll bring a mike to you. First question? [After last question] The buffet lunch is set up just outside this room. Please make your selections and come back in. Scott and I will be joining you. And please join us tomorrow right here, 12:15 PM, for a briefing featuring Oliver Ratzesberger of eBay and the Teradata Petabyte Power Players. Thank you for joining us today.
Teradata Unity Imad Birouty, Program Marketing Manager 7/19/2011
User Session Routing Great for session and workload control across systems, especially when high availability failover is required Individual Query Routing Great when systems contain different data and/or for fast setup where session control isn’t required
Databases objects don’t have to be identical (tables, users, view, macros,…)
Data Synchronization John 123 82.50 Bill 858 12.50 Jill 545 72.12 Tom 684 64.29 Joe 367 97.24 Bob 826 55.19 John 123 82.50 Bill 858 12.50 Jill 545 72.12 Tom 684 64.29 Joe 367 97.24 Bob 826 55.19 Tom 684 64.29 Joe 367 97.24 Bob 826 55.19
E D C Unity Teradata System B Teradata System A Query Routing Schema/Data Synchronization
SQL data updates
Sequencer Data Dictionary Parser Monitoring/Control Data Synchronization SQL Multicast Data Loads * Replication * Data Mover * Status * This method of data sync can coexist with Unity, but needs to be managed and coordinated outside Unity Users/ Applications/ SQL Utilities Users/ Applications/ SQL Utilities Users/ Applications/ SQL Utilities Teradata Multi-System Manager Monitoring Administration Operational Control