SAS Visual Analytics is not an in-memory database. In fact, it liberates customers (and SAS) from dependence on an expensive third-party database because it holds data in memory on a rack of blades running the Hadoop Distributed File System. Customers won't have to know anything about configuring or running Hadoop, said SAS, because all the deployment, provisioning, and administration will be handled by the platform's SAS LASR Analytic Server. The platform has been tested with more than 20,000 columns and 1 billion rows of data, according to SAS, and to scale out, customers simply add more nodes.
More Insights
Webcasts
- The Critical Importance of High Performance Data Integration for Big Data Analytics
- Mobile DevOps: Achieving continuous delivery with multiple front ends and complex backends in Banking, Financial Services, and Insurance
White Papers
More >>Reports
More >>SAS gave InformationWeek a preview of SAS Visual Analytics last month, describing it as a self-service business intelligence (BI) product. It does fit that description in that it supports fast query and reporting. But the true power and distinguishing characteristic of the platform is its ability to apply analytical computations to this massive pool of data held in memory.
"If you look at what SAS is doing versus others, we're not just exploring past activity, we're supporting analyses that are predictive, so people can see into the future of their business performance," Jim Davis, SAS' senior VP and chief marketing officer, told InformationWeek.
[ Want more on SAS data-integration capabilities? Read SAS Taps Hadoop For Big Data Analytics. ]
As an example, Davis said that predictive marketing campaign optimization efforts that now take eight to 10 hours in a conventional SAS environment can be completed in less than three minutes on the platform, and bank-risk calculations that formerly took 18 hours now take 15 minutes.
With SAS data-integration capabilities, the vendor can pull data into the Visual Analytics cluster from virtually any relational database or application data source. In addition to the LASR Analytic Server, platform components include the SAS Visual Analytics Explorer (VAE) data-visualization interface; a Designer for creating reports and dashboards; an admin interface for managing data, users, and security; and Visual Analytics Mobile, available initially for iPad, an app for viewing reports and downloading visualizations and supporting data from the LASR server.
In a VAE demonstration at SAS' New York office, a SAS executive dragged 10 variables onto the tool's palette and started using drag-and-drop filters, check boxes, and sliders to narrow down the data set. The tool automatically suggests the most appropriate visualization for the data used, but users can also manually choose from among options including bar, line, scatter plot, bubble, geographic, heat map, histogram, and box plot charts.
SAS intends to run "virtually all" of its vertical-industry and function-specific analytic applications on Visual Analytics, according to Davis. Starting today, SAS can run marketing automation and value- and risk-analysis apps for banking on the platform. Next up will be retail price optimization apps for determining the best everyday, promotional, and clearance pricing.
SAS Visual Analytics is appliance-like in that it runs on a SAS-specified configuration of commodity blade servers. Customers will buy the suggested hardware directly from vendors such as HP and Dell, and the software is downloaded, installed, and remotely configured by SAS. Reference configurations begin with an eight-blade server with 96 processor cores, 768 gigabytes memory, and 4.8 terabytes of disk storage. High-end reference configurations make use of 96 blades with 1,152 cores, 9.2 terabytes of memory, and 57.6 terabytes of disk storage.
With in-memory products gaining ground in recent years, the handwriting has been on the wall for older BI technologies, but Davis said VAE "shoots OLAP [online analytical processing] in the head for good" in terms of defining dimensions on the fly without having to build a cube--which can take hours or even days--in order to get rapid data interaction and analysis. That's a threat to displace widely deployed but aging Oracle Essbase and IBM Cognos deployments.
But SAS' strongest play with Visual Analytics is to supercharge its predictive analytics capabilities. The time shift from hours to minutes promises to enable customers to do more analysis with more data and greater accuracy than ever before.
Look for tension between customization and mass appeal as SaaS providers try to keep large customers happy while staying true to the multitenant model. Find out the whole story in our SaaS Adoption Soars, Yet Deployment Concerns Linger report. (Free registration required.)
