Belarus Big Data User Group Meetup #23

18.45 – 19.00

Welcome Coffee

19.00 – 19.50

«Apache Beam: unified API or next useless standard for dataprocessing» Alexey Demin (Java Developer)

A single standard API for data stream and batch. Not pertained to any specific product but the ability to run on Dataflow / Spark / Flink / Apex. Writing tasks both in Java, and in Python. The active joint development of the leading companies in this field. All this is a dream or the birth of a new common standard for data processing systems.

19.50 – 20.10

Coffee-break

20.10 – 21.00

«Dynamic and flexible resource scheduling for heterogenous workloads with Mesos and DC/OS» Alexander Rukletsov (Distributed Systems Engineer), Mesosphere

Operating apps at scale becomes the new norm. Containers, cluster managers, resource negotiators—these words pop up more and more often in blogs, technical reports, papers, and news. Some folks rely on these modern technologies in production, and many are in experimental phases. Efficient scheduling of heterogenous workloads can significantly cut costs and reduce efforts required to maintain and run server infrastructure. In this talk Alex will touch on prerequisites for success of containers and cluster managers, will give a technical overview of the scheduling part of Mesos, and wrap up with Spark on Mesos 101.

For whom: devops, cluster operators, data engineers, distributed systems engineers.