Wednesday 17 June 2015

NOTE: SAS Increasingly Embracing Virtuality & Cloud (and Hadoop) #SASGF15

Last week's UK SAS Forum encouraged me to dig-out my notes from this year's SAS Global Forum and type them up into some blog posts.

One of the notable trends at this year's SAS Global Forum was cloud and virtuality, both providing SAS sites with increased choice and flexibility.

The SAS University Edition has long been available free as a "vApp", i.e. an application packaged entirely into a virtual image that you can launch on your desktop (with the use of VMware Player or Oracle VirtualBox). This provides SAS Studio, Base SAS, SAS/STAT, SAS/IML and SAS/ACCESS (for PC file formats). The vApp approach means little or no installation effort, and the ability or un the same package on any operating system that supports VMware Player or Oracle VirtualBox, e.g. Windows and Mac OS.

At SGF, SAS announced the availability of University Edition via Amazon Web Services (AWS), meaning the power of SAS can now be accessed from any supported browser on any platform (including my ChromeBook!). SAS University Edition in AWS Marketplace is eligible for the AWS Free Tier program, which provides free access to first-time AWS subscribers for 12 months, up to 750 hours a month (subsequent AWS usage fees apply).

In vApp form, the package includes an HTTP sever, operating system, and SAS software, so it's a full 3-tier solution in a black box (meaning far fewer visits to the excellent SAS Community for Admin & Deployment!).

Another vApp, recently made available by SAS, is the Data Loader for Hadoop. Sadly, unlike the University Edition, this is free for a 90 day trial only, but if you are integrating your SAS platform with one or more Hadoop instances it's certainly worth a look. Data Loader for Hadoop provides self-service big data preparation, data quality and data integration for business analysts and data scientists. The point-and-click user interface enables users to prepare, integrate and cleanse big data faster and easier without writing code. Additionally, power users can run SAS code and data quality functions faster on Hadoop for improved productivity and reduced data movement. The most current release is SAS Data Loader for Hadoop 2.2.

If you're keen to earn more about Hadoop it's worth taking a look at the Hortonworks Sandbox. This too is a vApp (and it's free). If you have Data Loader for Hadoop alongside Hortonworks Sandbox then you'll find that Data Loader will configure itself against the cluster on first use, i.e. it will install the SAS embedded process within the cluster so that you can more efficiently execute your SAS work within the cluster.

There's plenty of information to help you get your SAS and Hortonworks technology working together. Look on the Hortonworks site (Explore the Possibilities of SAS Software and Hortonworks Sandbox) site and the SAS site (Step-by-Step Installation & Configuration of SAS Data Loader for Hadoop (Trial Edition) With Hortonworks Sandbox on Windows, and SAS Data Loader for Hadoop Video Tutorials).

Have fun!