Mark Smith's Analyst Perspectives

Teradata continues to expand its information management and analytics technology for big data to meet growing demand. My analysis last year discussed Teradata’s approach to big data in the context of its distributed computing and data architecture. I recently got an update on the company’s strategy and products at the annual Teradata analyst summit. Our big data analytics research finds that a broad approach to big data is wise: Three-quarters of organizations want analytics to access data from all sources and not just one specific to big data. This inclusive approach is what Teradata as designed its architectural and technological approach in managing the access, storage and use of data and analytics.

Teradata has advanced its data warehouse appliance and database technologies to unify in-memory and distributed computing with Hadoop, other databases and NoSQL in one architecture; this enables it to move to center stage of the big data market. Teradata Intelligent Memory provides optimal accessibility to data based on usage characteristics for DBAs, analysts and business users consuming data from Teradata’s Unified Data Architecture (UDA). Teradata also introduced QueryGrid technology, which virtualizes distributed access to and processing of data across many sources, including the Teradata range of appliances, Teradata Aster technology, Hadoop through its SQL-H, other databases including Oracle’s and data sources including the SAS, Perl, Python and even R languages. Teradata can provide push-down processing of getting data and analytics processed through parallel execution in its UDA including data from Hadoop. Teradata QueryGrid data virtualization layer can dynamically access data and compute analytics as needed making it versatile to meet a broadening scope of big data needs.

Teradata has embraced Hadoop through a strategic relationship with Hortonworks. Its commercial distribution, Teradata Open Distribution for Hadoop (TDH) 2.1, and originates from Hortonworks. It recently announced Teradata Portfolio for Hadoop 2, which has many components. There is also a new Teradata Appliance for Hadoop; this is its fourth-generation machine and includes previously integrated and configured software with the hardware and services. Teradata has embraced and integrated Hadoop into its UDA to ensure it is a unified part of its product portfolio that is essential as Hadoop is still maturing and is not ready to operate in a fully managed and scalable environment.

Teradata has enhanced its existing portfolio of workload-specific appliances. It includes the Integrated Big Data Platform 1700, which handles up to 234 petabytes, the Integrated Data Warehouses 2750 for up to 21 petabytes for scalable data warehousing and the 6750 for balanced active data warehousing. Each appliance is configured for enterprise-class needs, works in a multisystem environment and supports balancing and shifting of workloads with high availability and disaster recovery. They are available in a variety of ratios including disks, arrays and nodes, which makes them uniquely focused for enterprise use. The appliances run version 15 of the Teradata database with Teradata Intelligent Memory and interoperate through integrated workload management. In a virtual data warehouse the appliances can provide maximum compute power, capacity and concurrent user potential for heavy work such as connecting to Hadoop and Teradata Aster. UDA enables distributed management and operations of workload-specific platforms to use data assets efficiently. Teradata Unity now is more robust in moving and loading data, and Ecosystem Manager now supports monitoring of Aster and Hadoop systems across the entire range of data managed by Teradata.

Teradata is entering the market for legacy SAP applications with Teradata Analytics for SAP, which provides integration and data models across lines of business to use logical data from SAP applications more efficiently. Teradata acquired this product from a small company in last year; it uses an approach common among data integration technologies today and can make data readily available through new access points to SAP HANA. The product can help organizations that have not committed to SAP and its technology roadmap, which proposes using SAP HANA to streamline processing of data and analytics from business applications such as CRM and ERP. For others that are moving to SAP, Teradata Analytics for SAP can provide interim support for existing SAP applications.

Teradata continues to advance JavaScript Object Notation (JSON) integration for support of document-oriented databases that are schemaless and semistructured. JSON has become a critical tool as more applications need to store and access data efficiently. NoSQL databases have become more popular recently: 25 percent of organizations in our big data analytics research are using them today, 20 percent  plan to use them within two years, and another 23 percent are evaluating NoSQL. With this focus Teradata provides for its customers application and operational support beyond just supporting data for analytic purposes.

Teradata continues expansion of its Aster Discovery Platform to process analytics for discovery and exploration and also advances visualization and interactivity with analytics, which could encroach on partners that provide advanced analytics capabilities like discovery and exploration. Organizations looking for analytic discovery tools should consider this technology overlap. Teradata provides a broad and integrated big data platform and architecture with advanced resource management to process data and analytics efficiently. In addition it provides archiving, auditing and compliance support for enterprises. It can support a range of data refining tasks including fast data landing and staging, lower workload concurrency, and multistructured and file-based data.

Teradata efforts are also supported in what I call a big data or data warehouse as a service and is called Teradata Cloud. Its approach is can operate across and be accessed from a multitenant environment where it makes its portfolio of Teradata, Aster and Hadoop available in what they call cloud compute units. This can be used in a variety of cloud computing approaches including public, private, hybrid and for backup and discovery needs. It has gained brand name customers like BevMo and Netflix who have been public references on their support of Teradata Cloud. Utilizing this cloud computing approach eliminates the need for placing Teradata appliances in the data center while providing maximum value from the technology. Teradata advancements in cloud computing comes at a perfect time where our information optimization research finds that a quarter of organizations now prefer a cloud computing approach with eight percent prefer it to be hosted by a supplier in a specific private cloud approach.

What makes Teradata’s direction unique is moving beyond its own appliances to embrace the enterprise architecture and existing data sources; this makes it more inclusive in access than other big data approaches like those from Hadoop providers and in-memory approaches that focus more on themselves than their customers’ actual needs. Data architectures have become more complex with Hadoop, in-memory, NoSQL and appliances all in the mix. Teradata has gathered this broad range of database technology into a unified approach while integrating its products directly with those of other vendors. This inclusive approach is timely as organizations are changing how they make information available, and our information optimization benchmark research finds improving operational efficiency (for 67%) and gaining a competitive advantage (63%) to be the top two reasons for doing that. Teradata’s approach to big data helps broaden data architectures, which will help organizations in the long run. If you have not considered Teradata and its UDA and new QueryGrid technologies for your enterprise architecture, I recommend looking at them.

Regards,

Mark Smith

CEO & Chief Research Officer