Tech2Tech
2009 Teradata R&D Excellence Awards
Teradata continues to lead the data warehousing and business intelligence (BI) industry with a dedication to innovation in its products and services. The R&D Excellence Awards program recognizes Teradata associates for their innovative ideas and technical achievements that lead to significant advances and improvements for Teradata and its customers. Please join us in congratulating the Outstanding Award winners for 2009.
Ahmad Ghazal, Engineering manager, Teradata Database
Query Rewrite enhancements
This feature adds predicate simplification and redundant distinct removal to the Query Rewrite component of the Optimizer. Existing join elimination and satisfiability and transitive closure (SATTC) are enhanced. The improved rewrites provide significant performance enhancements to real-life problems. In particular, partition elimination on non-numeric types is made possible, and enhanced join elimination can help achieve vertical partitioning.
Grace Au, Software engineer, Teradata Database
Scalar subquery (SSQ)
This feature allows a subquery that returns a single value to be used anywhere an expression can be used in a query. The innovation not only provides a powerful tool for application developers to write complex queries but also allows migration of applications with SSQ to the Teradata Database.
Ken Nishihara, Software engineer, Teradata Database
Sorana Rabinovici, Software engineer, Teradata Database
Segment cache for messaging
Excessive root disk I/O on Linux and Windows 64 systems running highly concurrent workloads resulted in unpredictable performance. On Linux systems, for instance, data written to the root disk throttled performance, since waiting for I/O caused unnecessary delays and tied up the CPU. Locking BYNET node-level message data segments in memory results in significant reduction in superfluous I/O to the root disk, which improves performance.
Greg Milby, Software engineer, Teradata Database
Dynamic user-defined types to support the SAS in-database initiative
A single new user-defined data type, the VARIANT_TYPE, is defined during runtime and can be used to package parameters of different data types as a single data type. This innovation enables the quick porting of SAS functionality to the Teradata Database via user-defined functions.
Steve Hurlbut, Software engineer, Teradata Platform and Client
Teradata Viewpoint Rewind
This Teradata system time machine allows any authorized Teradata Viewpoint user to easily step back with the Rewind slider and view exactly what happened in the system at a given point in time. Among its benefits, this feature helps DBAs diagnose the root cause of a performance problem by going back to when it first occurred.
Chinh Kim Nguyen , Hardware engineer, Teradata Platform and Client
Curtis Stehley, Software engineer, Teradata Platform and Client
Full-section bandwidth network for Teradata data warehouse appliances
The innovation is taking advantage of Ethernet standard protocols such as 802.1q and 802.3ad, supported in all commodity Ethernet switches, to produce a full-section bandwidth topology for large Ethernet networks. Using low-cost Ethernet switches, a network of up to 552 ports can be produced. This allows an Ethernet-based version of the BYNET for all Teradata platforms.
Todd Papaioannou, Engineering fellow, Teradata R&D CTO
Teradata Viewpoint
A state-of-the-art Web 2.0 application-delivery framework, Teradata Viewpoint takes advantage of the open-source system to increase productivity and create opportunities via portlets. It also developed, based on field input, an incorporated graphical user interface. Teradata Viewpoint has transformed the way customers view Teradata management tools and the way client applications are delivered.
Ranjan Priyadarshi Software engineer, Teradata Database
GROUP BY and DISTINCT planning enhancements
This feature unifies GROUP BY and DISTINCT, thus removing performance discrepancy between them. In addition, all aggregation operations are incorporated into cost-based join planning, enhancing the Optimizer’s ability to pick optimal plans. The innovation significantly improves aggregation performance, which is key to data warehousing applications. It provides the infrastructure for continuing enhancements in the aggregation area.
Yi Xia, Software engineer, Teradata Database
Complex outer joins: design and enhancements
This project redesigns and incorporates major enhancements to the complex outer join logic that is prevalent in numerous customer query workloads. Illegal joins and plan failures that are major sources of wrong results are effectively prevented. The new design also improves performance of complex outer joins and makes it easier to maintain the code.