HDFS Design Principles and the Scale-out-Ability of Distributed Storage

The presentation will cover commonly discussed topics about reliability, availability, and scalability of the Hadoop Distributed File System. It will address

key design principles of HDFS as a distributed storage platform for Apache Hadoop,

author’s experience of building large Hadoop clusters,

competing approaches to HDFS HA,

scalability challenges for Hadoop, and

outline of one way out of HDFS scalability trap

Speaker: Konstantin V. Shvachko

Konstantin is a veteran Hadoop developer and an author of articles and presentations about Hadoop scalability. He has a practical experience in building large Hadoop installations at eBay and Yahoo. Konstantin specializes in efficient data structures and algo rithms for large-scale distributed storage systems. He is a committer and a member of the Apache Hadoop PMC.

----------------

FREE for SVForum members.

$20 for non-members.