Live-Stream Link: http://www.ustream.tv/channel/spark-meetup-feb-5-2014
Ali Ghodsi of Databricks will be presenting 2 talks at the Huawei offices in Santa Clara, CA on Wednesday Feb 5.
• TGF: Performing advanced analytics in Shark through Table Generating Functions
This meetup covers two new features, one for Shark and one for Spark. For Shark, we introduce Table Generating Functions (TGFs). These enable users to perform advanced analytics, such as calling ML libraries, from Shark. TGF is a flexible mechanism that lets you wrap existing Spark libraries, supplying them with parameters, and getting results back as tables. The mechanism builds on the new enhanced RDD and SQL table convertors available in Shark.
• SIMR: Seamlessly launching Spark jobs on MapReduce v1 clusters
For Spark, we now support the ability to launch Spark jobs on Hadoop MapReduce v1 clusters through SIMR (Spark In MapReduce). This deployment mode for Spark is very seamless as it only requires downloading three files and access to an MR1 cluster. SIMR also supports running the Spark REPL inside MR clusters.
This Meetup will be live streamed and later added to YouTube.