Oracle had introduced Exalytics last October at Oracle OpenWorld to go with its Exalogic cloud-systems server and Exadata storage server.
Exalytics, Oracle's new in-memory big data analytics appliance that moves "at the speed of thought" (CEO and co-founder Larry Ellison's words), on Feb. 27 graduated from theory to general availability (GA) status in the Oracle product catalog.
Oracle had introduced Exalytics last October at Oracle OpenWorld to go with its Exalogic cloud-systems server and its Exadata storage server. After having been tested for several months with an undisclosed number of customers, Exalytics now has been sanctioned as a ready-for-prime-time product.
Exalytics is a fast-moving DRAM (dynamic memory) and NAND flash-based, large workload-type analytics server that runs everything in memory on its own software or that of batch analyzers Apache Hadoop or EMC Greenplum. Spinning disks, be gone.
Data Travels at the 'Speed of Thought'
"This really is what we consider 'instantaneous in-memory response,'" Ellison said about the appliance in his OpenWorld keynote Oct. 2. "You cannot get any faster than this. The Essbase analytics are tuned specifically for in-memory processing. These things work basically at the speed of thought."
Oracle claims the Exalytics appliance can scan and analyze 1 billion or more records at a time.
The appliance is really a fast-to-begin-with Oracle Sun server that's been reconfigured with all solid-state memory and storage. It's composed of a Sun Fire X4470 M2 server, which is a four-socket, 3U-size box running Intel's multicore (in this case, 10 CPUs) Westmere-EX Xeon E7-4800 chips. As such, it features 40 processors. Each box can store an impressive amount of 5TB to 10TB of deduplicated, compressed data in memory.
Now that is a serious pile of non-spinning-disk storage.
The real secret sauce, however, begins with its solid-state brains: a full 1TB of DDR3 main memory and six more flash disk drives packed onto the chassis. It features 40G-bps Infiniband connectivity with optional 10GbE ports. That is a serious amount of dynamic memory.
Oracle's parallelized TimesTen relational online transaction processing (OLTP) and Essbase parallel online analytical processing (OLAP) databases are what process the Big Data workloads. Oracle recently updated TimesTen for the first time in three years.
The TimesTen In-Memory Database for Exalytics also is available for Oracle's Exalytics In-Memory Machine as an application-tier in-memory database cache. This new release enables Oracle Business Intelligence Foundation Suite, which comes with it, a 20-times faster response time improvement for visualization and up to five times more data to be stored in memory with advanced columnar compression capabilities, Oracle Product Manager Gaurav Rewari told eWEEK.
On the management side, Exalytics has something called "heuristic adaptive in-memory caching" that calculates what data is best stored in-memory for overall best performance, and adapts to changing workloads, Rewari said. Unlike other appliances or in-memory tools, Oracle Exalytics does not constrain applications only to data that fits in-memory, but also leverages connected databases, data warehouses and OLAP sources.
More than 80 Oracle BI and Oracle Hyperion performance management applications are currently available for use with Oracle Exalytics -- and no application changes are required. These include all horizontal Oracle BI applications, such as Oracle Financial Analytics, Oracle Human Resources Analytics, Oracle Sales Analytics, and Oracle Procurement and Spend Analytics, among others; and Oracle Hyperion Planning, for planning and forecasting.
Supports Most Enterprise Data Sources
Because Exalytics is an open-standards system, it supports Oracle and a high number of third-party enterprise data sources in heterogeneous IT environments, Rewari said. It can access and analyze data from Oracle or third-party relational, OLAP or other data sources including IBM DB2, IBM Netezza, Microsoft SQL Server and Analysis Services, SAP Business Information Warehouse (BW), Sybase Adaptive Server Enterprise (ASE) and Teradata Warehouse, among others, in any combination.
The Exalytics hardware lists at $135,000, but that's not taking into consideration software licensing, service and maintenance SLAs. Go here to download a whitepaper on Exalytics (PDF).
Chris Preimesberger was named Editor-in-Chief of Features & Analysis at eWEEK in November 2011. Previously he served eWEEK as Senior Writer, covering a range of IT sectors that include data center systems, cloud computing, storage, virtualization, green IT, e-discovery and IT governance. His blog, Storage Station, is considered a go-to information source. Chris won a national Folio Award for magazine writing in November 2011 for a cover story on Salesforce.com and CEO-founder Marc Benioff, and he has served as a judge for the SIIA Codie Awards since 2005. In previous IT journalism, Chris was a founding editor of both IT Manager's Journal and DevX.com and was managing editor of Software Development magazine. His diverse resume also includes: sportswriter for the Los Angeles Daily News, covering NCAA and NBA basketball, television critic for the Palo Alto Times Tribune, and Sports Information Director at Stanford University. He has served as a correspondent for The Associated Press, covering Stanford and NCAA tournament basketball, since 1983. He has covered a number of major events, including the 1984 Democratic National Convention, a Presidential press conference at the White House in 1993, the Emmy Awards (three times), two Rose Bowls, the Fiesta Bowl, several NCAA men's and women's basketball tournaments, a Formula One Grand Prix auto race, a heavyweight boxing championship bout (Ali vs. Spinks, 1978), and the 1985 Super Bowl. A 1975 graduate of Pepperdine University in Malibu, Calif., Chris has won more than a dozen regional and national awards for his work. He and his wife, Rebecca, have four children and reside in Redwood City, Calif.Follow on Twitter: editingwhiz