This announcement is a very exciting prospect for some but may strike fear into others. In my blog, I will entertain some of the interesting prospects of bringing together these technologies. I also hope to allay some fears as well.
One of the biggest announcements at Teradata Partners 2015 is that Aster will run on Hadoop. Many of our customers have already invested in a Hadoop data lake and want to do more than just store data. Storing data is helpful but not all that interesting. What if you could easily do advanced analytics without having to move data out of the lake? What if you had the power of Aster’s multi-genre analytics running on Hadoop? This is exactly what Aster Analytics on Hadoop is all about.
Almost every day I hear about a new Hadoop project or offering. That means a new approach, a new tool to learn, and usually a lot of programming. With Aster, you have a variety of advanced analytics at your fingertips, ready to take advantage of your data lake. With Aster and its plug-and-play SNAP framework, analysts and data scientists can use a variety of analytics delivered through a common optimizer, executor, and unified interface. Aster offers many different genres of analytics: ANSI SQL, Machine Learning, Text, Graph, Statistics, Time Series, and Path, and Pattern Analysis. Aster on Hadoop is a big win for data scientists, as well as for people who already know and love Aster.
For those who know Aster, Aster on Hadoop might sound daunting, but don’t fret. Everything works the same. You have the same statement interface ‘SELECT * FROM nPath…’ You still have ACT, ncluster_loader, ncluster_export, and Aster Management Console. You can still run ANSI SQL queries and connect to disparate data sources through QueryGrid and SQL-H. AppCenter allows anyone to perform advanced analytics using a simple web interface. Aster Development Environment enables programmers to build their own custom SQL-MR and SQL-GR functions. In other words, everything works the same. The only difference is that it is all running inside Hadoop, enabling a whole new group of people to participate in the Hadoop experience. If you have made a large investment in Hadoop and want to exploit the data located there, then Aster on Hadoop is for you.
One of the biggest complaints I hear from clients is, “We built a data lake and we want to do analytics, but it’s too hard.” Aster is adaptable to your Hadoop environment and the data you’ve landed there. Aster on Hadoop also means no new appliance; no need to find room in your data center to park a new rack of Aster. There’s no data movement across platforms or across the network; you process data right where it is. Aster on Hadoop runs natively inside Hadoop so you have access to HDFS file formats and a variety of connectors to other JDBC/ODBC compliant data sources. Staff who know ANSI SQL are perfectly positioned to use Aster on Hadoop, and with a little training, they’ll be performing advanced analytics in no time.
Organizations have made huge strides and investments in their Hadoop ecosystem and many are using it as a repository for big data, but that’s not enough. Organizations rightly want to exploit the data contained in Hadoop to gain new insights. Today Aster is being used to solve real world business problems through its multi-genre analytic capabilities. Aster on Hadoop will lower the barriers to entry. It’s a big step in realizing real business value from Hadoop and finally achieving a positive ROI. If you’re an existing Aster client, there’s no need to worry: it all works the same. Teradata Aster on Hadoop democratizes analytics and brings solution freedom to Hadoop! It’s Hadoop for the rest of us.