Presented By O'Reilly and Cloudera
Make Data Work
March 13–14, 2017: Training
March 14–16, 2017: Tutorials & Conference
San Jose, CA

Spark at scale in Bing: Use cases and lessons learned

4:20pm5:00pm Wednesday, March 15, 2017
Spark & beyond
Location: LL21 C/D Level: Beginner
Secondary topics:  Architecture, Data Platform, Media
Average rating: ***..
(3.00, 3 ratings)

Who is this presentation for?

  • Software developers, data engineers, and data architects

Prerequisite knowledge

  • A basic working knowledge of Spark and Kafka

What you'll learn

  • Understand Spark use cases and architecture
  • Learn how to deal with scale issues in Spark applications (for example, when processing messages from a very large number of Kafka topics and partitions or checkpointing massive state)

Description

Apache Spark plays a key role in addressing several big data challenges in Bing. The diverse set of capabilities in Spark enables a variety of internet-scale workloads that power Bing services. The value Spark adds to the business and how well it fits with the existing data platform architecture complementing existing internal and external big data frameworks is clearly the driver behind the adoption of Spark for various next-gen data processing investments in Bing.

Kaarthik Sivashanmugam shares the Bing team’s experiences with Spark, discussing how Spark is employed in the use cases and covering batch processing of document corpus spanning the web and near real-time processing of events corresponding to hundreds of millions of search queries. Kaarthik also explores the challenges the team faced in adopting Spark and implementing scalable data processing pipelines and explains how they influenced the team in customizing Spark and building extensions.

Photo of Kaarthik Sivashanmugam

Kaarthik Sivashanmugam

Microsoft

Kaarthik is a Principal Software Engineering Manager in the AI Platform group at Microsoft. In his current role, he is leading a team of software engineers and applied scientists in implementing large scale training workloads on Azure Machine Learning service and enhancing the service to make it the best cloud-platform for data scientists and ML engineers. Prior to this work, Kaarthik was involved in the development of near real time data processing platform and GPU infrastructure for deep learning.