Presented By O'Reilly and Cloudera
Make Data Work
Dec 4–5, 2017: Training
Dec 5–7, 2017: Tutorials & Conference

Apache Kylin: Advanced tuning and best practices with KyBot

Dong Li (Kyligence), Luke Han (Kyligence)
2:35pm3:15pm Thursday, December 7, 2017
Average rating: *****
(5.00, 1 rating)

Who is this presentation for?

  • Anyone working with Apache Kylin

Prerequisite knowledge

  • Familiarity with Hadoop, OLAP, and Apache Kylin

What you'll learn

  • Understand optimization rules and best practices for Apache Kylin


Apache Kylin is an extreme distributed OLAP engine on Hadoop. The basic idea of Apache Kylin is replacing online calculation with offline precalculation, and its most important concept is cube. Apache Kylin provides a theory of cube designing with multiple optimization rules. Well-tuned cubes bring about the best performance with the least cost but require a comprehensive understanding of tuning principles to use. Tuning is the key to keep performance no matter how business requirements change, and leveraging AI into big data tuning can save lots of workforce and learning costs. The solution for Kylin is KyBot.

Dong Li and Luke Han explain advanced tuning and introduce KyBot, an online tuning service for Apache Kylin that provides self-service functionalities for system diagnostics and tuning. Based on AI algorithms over historical logs analysis and knowledge base built by Apache Kylin core developers, KyBot provides rich lifecycle tracking dashboards for queries and building jobs profiling, which help to find bottlenecks in seconds. More to the point, KyBot also provides suggestions about tuning and optimization to download, with which users can simply done tuning by directly importing what’s downloaded.

Photo of Dong Li

Dong Li


Dong Li is a PMC Member and Committer of Apache Kylin, and Founding Member and Head of Product at Kyligence, focus on technology research and product strategy around AI/Big Data/Cloud ecosystem. Previously, Dong was a senior software engineer at eBay and a software development engineer at Microsoft.

Photo of Luke Han

Luke Han


Luke (Qing) Han is the coounder and CEO of Kyligence, which provides a leading intelligent data platform powered by Apache Kylin to simplify big data analytics from on-premises to the cloud. Luke is the cocreator and PMC chair of Apache Kylin, where he contributes his passion to driving the project’s strategy, roadmap, and product design. For the past few years, Luke has been working on growing Apache Kylin’s community, building its ecosystem, and extending its adoption globally. Previously, he was big data product lead at eBay, where he managed Apache Kylin, engaged customers, and coordinated various teams from different geographical locations, and chief consultant at Actuate China.