AT ORACLE: News
Go Big and Go FastBy Fred Sandsmark
New Oracle engineered systems deliver big data and high-speed visual analytics.
According to Gartner, big data, next-generation analytics, and in-memory computing are three of the top strategic technologies for 2012. To help enterprises stay ahead of the curve, Oracle introduced two new systems at Oracle OpenWorld 2011 that focus strategic technologies on some of today’s most daunting information processing challenges. The first product, Oracle Big Data Appliance, is built to acquire, organize, and load large volumes of unstructured data; the second, Oracle Exalytics In-Memory Machine, uses in-memory database software and an optimized business intelligence platform to provide extremely fast visual data analytics.
Oracle Big Data Appliance
Andy Mendelsohn, senior vice president of database server technologies at Oracle, introduced Oracle Big Data Appliance during his Oracle OpenWorld general session on big data. Mendelsohn began his presentation by describing three characteristics of big data: massive volumes, high velocity or frequency, and datasources that are varied and often unstructured (such as e-mail, sensor data, and smart meter readings).
Oracle Big Data Appliance is engineered to address those characteristics: it acquires unstructured data, organizes and filters it to discover the most-valuable nuggets of information, and loads the result into a data warehouse for analysis and decision-making. Its hardware is a robust rack of 18 Sun Fire X4270 M2 servers totaling 216 processor cores, 864 GB of memory, and 432 TB of storage. InfiniBand interconnects are used to integrate these components and connect Oracle Big Data Appliance with Oracle Exadata Database Machine and Oracle Exalytics In-Memory Machine.
The software included with—and tuned for—Oracle Big Data Appliance includes an open source distribution of Apache Hadoop, for handling data-intensive applications; Oracle NoSQL Database, Enterprise Edition, a scalable key-value database; Oracle Data Integrator Application Adapter for Hadoop, which simplifies data integration from Hadoop; and Oracle Loader for Hadoop, which provides an easy way to load data from a Hadoop cluster into Oracle Database or Oracle Exadata Database Machine.
With these hardware and software components, engineered to work together, Oracle Big Data Appliance provides a complete solution for handling big data’s volume, velocity, and variety, and helps companies derive value from that data, Mendelsohn said. “Oracle is the first vendor to offer customers a complete and integrated set of products to address critical big data requirements, unlock efficiencies, simplify management, and create data insights that maximize business value,” he explained.
Oracle Exalytics In-Memory Machine
In his opening keynote at Oracle OpenWorld, CEO Larry Ellison introduced Oracle Exalytics In-Memory Machine, the industry’s first engineered system for analytics. Oracle Exalytics combines Oracle’s Sun hardware; an optimized version of Oracle Business Intelligence foundation with enhanced data visualization and performance capabilities; and a version of Oracle TimesTen In-Memory Database optimized to run business analytics. Oracle Exalytics can access data from any Oracle or non-Oracle relational, online analytical processing, or unstructured datasource and provide visual analytics at the speed of thought. “There’s nothing faster than this thing,” Ellison said.
Oracle Exalytics is built on an Oracle’s Sun Fire server with Intel Xeon E7-4800–series processors totaling 40 cores, coupled with a full terabyte of RAM. The software stack is tuned for the hardware and optimized for business intelligence. A user interface provides interactive data visualization, enabling even casual users to explore large data sets and spot patterns, trends, and outliers.
Oracle Exalytics uses adaptive in-memory technology to ensure that the most-important data sets are always in memory and automatically tunes cached data by constantly monitoring user queries and moving most-used data from storage into memory.
Based on Oracle’s internal testing scenarios, Oracle Exalytics provides more than 20 times faster response times in comparison to conventional BI software running on generic hardware configurations—a difference that Ellison believes will change the practice of analytics. “When you have a 20-times improvement, you operate differently,” he said. “You ask more questions. You get better answers.”
Fred Sandsmark is a regular contributor to Oracle Magazine.