Teradata Columnar September 2011 Copyright 2011 Teradata Corporation

Teradata Columnar Teradata is the hybrid row and column database for the enterprise • Extreme performance > Improves data warehouse performance • Dramatic compression > Automatic compression of columnar data • Ease of Use > Simple to use, automated, dynamic • Integrated > Full Integrated Data Warehouse capabilities with columnar > Use in IDW or for analytical data marts September 2011 Copyright 2011 Teradata Corporation 2

What is Teradata Columnar? • Description > Teradata Columnar is a new physical database design technique that allows tables to be hybrid horizontal row and/or vertical column partitions • Benefits > Performance – Improves query performance via column partition elimination by eliminating need to access all the data in a row table > Compression – Through a new Auto-Compression capability data is compressed automatically September 2011 Copyright 2011 Teradata Corporation 3

Teradata Columnar Cust. Number Name Address City State Zip For this table … In row format, each unit of storage (block) has all of the values for a row(s) In column format, each unit of storage (block) has values for a single column September 2011 Copyright 2011 Teradata Corporation 4

Teradata Columnar Who are all the customers who live in San Diego? Cust. Number City In row format, database reads all data to scan City column spread over all blocks with other data Cust. Number City San Diego September 2011 Copyright 2011 Teradata Corporation In column format, database only reads City data and Cust. Number data because data in blocks is for one column 5

Teradata Columnar: Extreme Performance • Improves data warehouse performance > Faster application response > Get more work out of the system > Reduces I/O for better performance • Column and row partition elimination • Hybrid row/column > True columnar storage rather than just compression mechanism > Use most efficient storage for data and access pattern • Optimizer chooses best query plan > Chooses between row and column data when both available with base table and Join Index > Takes advantage of all existing Teradata performance features September 2011 Copyright 2011 Teradata Corporation 6

Teradata Columnar: Dramatic Compression • Compression > Store more data > Reduced I/O for better performance > Store more data in fast SSD drives • Multiple compression mechanisms automatically selected • Columnar auto-compression and existing user specified compression mechanisms on same data > MVC, algorithmic, block-level September 2011 Copyright 2011 Teradata Corporation 7

Teradata Columnar: Ease of Use • Automated for maximum ease of use and simplicity > Compression mechanism chosen automatically > Values for dictionary compression chosen automatically > Dynamic compression; re-evaluated for each “container” > Reduced design and analysis time > Minimal administration effort • Hybrid Row and Column > Teradata Columnar automatically determines if column should be row or column format • Ease of management > Reduction of DBA work on aggregates or PI maintenance > SELECT queries do not change with columnar data September 2011 Copyright 2011 Teradata Corporation 8

Teradata Columnar: Integrated • Advanced Tuning > Row Partitioning, Secondary indexes, Join indexes • Rich Optimizer > Rich SQL language, Teradata optimizer • Mission Critical Availability > 100% Performance Continuity, Full redundancy, Failover • Workload Management > Teradata Active System Management • Hybrid Storage > Teradata multi-temperature support (TVS) • Scalability > Multidimensional Scalability to 100 Terabytes and Petabytes • In-Database Analytics > In database processing September 2011 Copyright 2011 Teradata Corporation 9

Teradata Columnar Messaging Teradata is the hybrid row and column database for the enterprise • Extreme performance > Improves data warehouse performance • Dramatic compression > Automatic compression of columnar data • Ease of Use > Simple to use, automated, dynamic • Integrated > Full Integrated Data Warehouse capabilities with columnar > Use in IDW or for analytical data marts September 2011 Copyright 2011 Teradata Corporation 10