May 21 · 6:30 PM
Bronto Software, Inc.
SAS and Hadoop - Chalk and Cheese? Or Fit for Each Other?
Speaker: Paul Kent, Vice President, Big Data at SAS
This talk describes our work at SAS as we shape our software to be a good citizen in the Hadoop eco-system. A look into the technical challenges of interfacing broadly with Hadoop and its subsystems: Hive, HiverServer2 and Kerberos, MapReduce/YARN, and persistent daemons like Impala.
We'll discuss which techniques work for us, and at what cost.
About the author:
Paul Kent is Vice President of Big Data initiatives at SAS. He spends his time between Customers, Partners and the Research & Development teams discussing, evangelizing and developing software at the confluence of big data and high performance computing. A datacenter rack full of current-generation 64bit x86 processors represents a very large aggregate memory space, thousands of threads and plentiful IO that can be harnessed
to solve problems at a much larger scale than we have traditionally been accustomed to.
Paul was previously Vice President of the Platform R&D division at SAS and lead groups responsible for the SAS foundation and mid-tier technologies -- Teams that develop, maintain and test Base SAS®, as well as related data access, storage, management, presentation, connectivity and middleware software products.