Teradata pioneers new Teradata Aster SNAP Framework

Data scientists and analysts have been plagued by a piecemeal big data strategy, rudimentary tools, lack of scalability and the use of desperate tools, technologies and processes. “Recent IDC research revealed that only 10 percent of organisations have the features and functionality needed to explore data and discover insights,” said Dan Vesset, program vice president, Business Analytics and Big Data, IDC. “There are clearly many opportunities for improvement. An integrated data discovery platform should be part of every organisation’s portfolio, for without this capability they have a real void in their business analytics strategy.”

“The innovative design of the Aster Discovery Platform will liberate data scientists around the world by reducing complexity, breaking down analytic silos and magnifying analytic ability,” said Scott Gnau, president, Teradata Labs. “The Aster Discovery Platform is the first to offer cross-analytic engine optimisation and data processing innovations that make it faster and easier for a wider group of users to leverage Aster’s analytics. The Aster Discovery Platform will fuel innovation and profoundly impact business results.”

The innovations in the Aster Discovery Platform include Teradata Aster SQL-GR™, a graph engine and the Teradata SNAP Framework™ (Teradata Aster Seamless Network Analytic Processing Framework™).

New Graph Analytic Engine
Users may customise their Aster analytics with Teradata Aster SQL, Teradata Aster SQL-MapReduce®, or the newly-developed Teradata Aster SQL-GR™ engine. These engines empower organisations to analyse data in new and interesting ways to uncover never-seen- before business insights.

The Aster SQL-GR graph engine is unique because it is scalable, processes big data in parallel and is not limited by system memory. Its analytic capability can extend to millions of nodes or processing units. The Aster SQL-GR engine enables native processing of large-scale analytic graph queries and pre-built graph functions and can be used for customer churn, product affinity, fraud detection and recommendation engines. For example, telecommunications providers can use graph analysis to look for high-traffic connections among selected users, which provides clues to fraudulent calling activity.

Teradata SNAP Framework™
The only solution of its kind, the Teradata SNAP Framework™ enables multiple analytic engines and file stores to be seamlessly “snapped” together based on the customer’s tailored discovery needs. The tightly-integrated components within the Teradata SNAP Framework provide data scientists with unmatched, next-generation analytic power and speed to insight through a seamless discovery workflow, which even a business analyst can use.
Any user is able to delve deeply into data for new competitive insights by leveraging multiple analytical capabilities – like graph, MapReduce, text, statistical, time series and SQL-based analytics. Business analysts only need to submit a single query in SQL or through a business intelligence tool and the new query executer and optimiser seamlessly engage a combination of analytic engines on petabytes of data from new and existing files stores for high performance processing at scale.

The Aster Discovery Platform enables rapid exploration and discovery from all data with next-generation analytics, which creates new, differentiated business insights. Any SQL-savvy analyst or business user will find it easy to use, yet it is powerful enough for the most sophisticated data scientists. The Teradata SNAP Framework, which integrates analytic engines and data stores, delivers an order of magnitude improvement in ease-of-use, analytic power and performance.

The Teradata SNAP Framework seamlessly and simultaneously integrates the engines and data stores, and then executes and optimises the query a cross analytic engines and data stores. The innovative Teradata Aster Integrated Optimizer understands the discovery query’s request and automatically runs the query utilising the most efficient use of resources, which drives the highest performance data processing at scale. The new Teradata Aster Integrated Executor orchestrates all types of queries by enabling communication among analytic engines.

An additional component that differentiates the Teradata SNAP Framework is the Teradata Aster Common Storage System and Services. It provides a set of common services designed to efficiently write data to the disk, and offer data replication, and snapshots. These features support a unified security model, high availability, reliability and manageability.

“We use the Teradata Aster Discovery Platform to better understand our customers’ unique needs. The resulting insights help us update the Gilt Group website with the most relevant, personalized product offers,” said Geoff Guerdat, director of data engineering, Gilt Group. “The Teradata Aster Discovery Platform 6 is a big step forward and the Teradata SQL-GR analytic engine and file store will offer real business value.”

The Aster Discovery Platform offers customers technical innovations unavailable from other vendors. Other vendors have stand-alone technology that can independently run SQL, MapReduce, text analytics, statistical analytics, or graph analytics. However, only Teradata has integrated them into a single solution with the power to invoke one or all of these analytic capabilities in an easy-to-use iterative SQL statement.

The new capabilities will accelerate the discovery process and empower business users to visualise information in exciting and valuable ways. For example, retailers search for specific low profit margin food items that influence the purchase of high profit items. The purchase of low margin salad dressings can lead to additional items that the shopper needs for a meal, such as the higher margin meats and wine. With this discovery, retailers are able to understand bridge products like olives and cheese and strategically locate items in the store and offer coupons to maximise margin.

"At Tableau, we are committed to making data accessible to everyone," said Dan Jewett, vice president, Product Management, Tableau Software. "We are excited about the big data analysis and discovery capabilities the latest release of Teradata Aster Discovery Platform brings to market. We believe our joint customers will benefit tremendously from the Platform's innovations like the SNAP Framework, unified SQL interface, a new graph analytic engine, and data stores.”

Teradata Aster File Store
Teradata Aster’s storage architecture provides storage options that are built for specific storage and analytic demands of data discovery. Users can select the new Teradata Aster File Store™ for all data, in addition to the existing data stored in row or column format. The Teradata Aster File Store is designed to quickly take in and store petabytes of raw data, provide storage management, and then make it readily available for preprocessing.
The Teradata Aster File Store is compatible with Apache® Hadoop™ and the Hadoop Distributed File System. This compatibility provides a high-speed connector between Hadoop and Aster offering seamless integration within the Teradata Unified Data Architecture™. The Teradata Aster File Store enables any file format in Hadoop to work natively.

Expanded Analytic Functions
Teradata has added new pre-built functions that can be accessed from a common SQL framework. Aster Discovery Platform also extends in-database analytics and discovery with the integration of Fuzzy Logix’s™ 600-plus advanced analytic algorithms. The analytic functions further empower the data scientist and analyst to rapidly dig into data with the insightful, interactive visualisations delivered through a Web browser and common business intelligence tools.

The Teradata Aster Discovery Platform runs on the Teradata Aster Big Analytics Appliance and integrates the Teradata Aster Database, SQL-MapReduce, and Apache Hadoop. The Aster Discovery Platform offers a pre-built analytics library of more than 80 SQL-MapReduce functions. It runs on proven Teradata platform hardware, SUSE® Linux operating system, and enterprise-class storage – all preinstalled into a power-efficient unit. The system can be live and up and running in just a few hours for rapid time-to-value.