BIO:
I'm Alexey Filanovskiy - part of the development organization in Oracle. I use to be Oracle Developer, but since 2010 Hadoop world captivated me with his head.
Since 2012 I'm working with Big Data technologies in Oracle. I've started my carrier in pre-sales organization and in 2015 moved in Dev Org.
In my current role I'm working with many customers all around the world and help them to be successful with their Big Data projects.
So, I'm kind of bridge between development and fields, bring and create best practices from Development to the field and bring feedback from the fields to Development for adjust roadmap of Oracle Big Data products.

Proposal:

Building High performance and Highly available Hadoop cluster. Practical lessons

In my current position in Oracle I'm working with many customers all around the world and help them to be successful with their Big Data projects.
So, I'm kind of bridge between development and fields, bring and create best practices from Development to the field and bring feedback from the fields to Development for adjust roadmap of Oracle Big Data products.
After accumulating all practical experience and gather feedback from customers we found really good default configs for Hadoop Clusters, which could fin for 90% of the customers for their needs.
It include 5 pillars:
- YARN performance tuning
- Spark performance tuning
- Set up Static Service pools
- Set up Dynamic Service pools
- Setup Hadoop cluster in Highly available way
In my speech I'd like to talk about key  aspects of well configured Hadoop clusters