Extreme OLAP Engine for Big Data

An open source distributed Analytical Engine from eBay to provide SQL interface and multi-dimensional analysis (OLAP) on Hadoop to support TB to PB size analysis.

• Hadoop: 2.2.0.2.0.6.0-61 or above
• Hive: 0.12.0.2.0.6.0-61 or above
• HBase: 0.96.0.2.0.6.0-61-hadoop2

Tested with Hortonworks HDP 2.1.3 and Cloudera Quickstart VM 5.1.

• Command hadoop, hive, hbase is workable on your hadoop cluster
• JDK Runtime: JDK7 (OpenJDK or Oracle JDK)
• Maven
• Git
• Tomcat (CATALINA_HOME being set)
• Npm

• For the quick tutorial, we assume that your hadoop has priviledges disabled which means any user could run hadoop and hive command

It is very easy to install Kylin for exploration/development. There are 3 supported options:

1. Quick Start (Docker)
2. Sandbox (HDP or CDH sandbox)
3. Dev Environment (IDE + Sandbox)

This is the easiest option to try Kylin. We have created docker container for Kylin (along with Hadoop, HBase and Hive) with help of SequenceIQ. The only thing you will need to do is to pull the container from the official Docker repository to be up and running in few minutes.

Features:

• Size - Light weight compared to downloading and setting up HDP or CDH sandbox.
• Cluster support - Supports multi node installation.
• Fully Automated - No manual steps. One command does it all

For a hands-on tutorial please visit Kylin Docker installation.

If you are free to install Kylin on your hadoop CLI machine or Hadoop sandbox, this is the most convenient scenario, for it puts everything in a single machine.

For normal users, the application in the above picture means Kylin Web, which contains a web interface for cube building, querying and all sorts of management. Kylin Web launches a query engine for querying and a cube build engine for building cubes. These two engines interact with the components in Hadoop CLI, like hive and hbase.

For a hands-on tutorial please visit On-Hadoop-CLI installation.

This is typically for development environment setup.

Applications does not necessarily mean Kylin Web, it can also be any test cases or java applications running at your local IDE(query engine and cube build engine can be launched programmatically in your code). In this case, it is no longer reasonable to assume that all the Kylin components reside in the same machine as your Hadoop CLI. Fortunately, Kylin still works under such condition with proper settings on your CLI machine.

For a hands-on tutorial please visit Off-Hadoop-CLI installation

• Web Site: http://kylin.io

• Google Group: Kylin OLAP Group

• Developer Mail: [email protected]

• How To Contribute: See wiki

• Presentation: Kylin Hadoop OLAP Engine v1.0

• Apache Proposal: Apache Kylin