Apache Spark has become a very hot topic in the world of Big Data Analytics. It is seen as an alternative to Map/Reduce with superior performance due to more in-memory processing, and as a means to perform analytics that combine a wide set of data sources. IDUG had a one-day track of Spark presentations at IDUG EMEA 2015 in Dublin, where this session was recorded. This video by IBM’s Eberhard Hechler is a excellent introduction to what Spark is, why it is of interest and how it works with DB2 for z/OS, as well as other data sources on z/OS.
Apache Spark is becoming increasingly important in the context of z Analytics. To expose z data from different subsystems, such as DB2 for z/OS, IMS, VSAM, etc., into a Spark environment, represents an opportunity to apply Spark analytics to z data sources, and to integrate analytical insight derived via Spark from other heterogeneous data sources. This presentation will describe the technical integration of the key Spark components (Spark Core, Spark MLlib) into a z Analytics environment with DB2 for z/OS and other subsystems. Also learn about possible use case scenarios that are derived from client engagements.