Scalable stream processing with Apache Kafka and Apache Samza

I have given this talk (with minor variations) at the following venues:

• ApacheCon EU, Budapest, Hungary, 18 November 2014. http://apacheconeu2014.sched.org/event/3633e195715f88c3357749d57b7b3b8c
• Unified Log London Meetup, London, UK, 2 December 2014. http://www.meetup.com/unified-log-london/events/218025352/
• Jfokus, Stockholm, Sweden, 4 February 2015. https://martin.kleppmann.com/2015/02/04/samza-at-jfokus.html

Abstract:

Samza, an Apache Incubator project, is a framework for processing and analysing high-volume data streams. It is built upon Apache Kafka and YARN (Hadoop 2.0). You can think of Samza as a real-time, continuously running version of MapReduce.

In this talk, Martin will show why stream processing is becoming an important part of the architecture of data-intensive applications, alongside storage and batch processing. We will explore how Samza works, and show how it reliably processes millions of messages per second. We will also examine what kinds of applications would benefit from using Samza.

Transcript

14. { eventType: PageViewEvent, timestamp: 1413215518, viewerId: 1234, sessionId: fa1afe101234deadbeef, pageKey: profile-view, viewedProfileId: 4321, trackingKey: invitation-email, …metadata about displayed content… }
23. { eventType: PageViewEvent, timestamp: 1413215518, viewerId: 1234, sessionId: fa1afe101234deadbeef, pageKey: profile-view, viewedProfileId: 4321, trackingKey: invitation-email, …metadata about displayed content… }
32. key = urn:linkedin:profile:1234 value = { eventType: ProfileEditEvent, timestamp: 1413215518, profile: { location: “Cambridge, UK”, industry: “Software”, positions: [ {job_title: “Author”, company: “O’Reilly”}, … ]}}
46. key = urn:linkedin:profile:1234 value = { eventType: ProfileEditEvent, timestamp: 1413215518, profile: { location: “Cambridge, UK”, industry: “Software”, positions: [ {job_title: “Author”, company: “O’Reilly”}, … ]}}
56. References (fun stuff to read) 1.  Martin Kleppmann: “Designing data-intensive applications.” O’Reilly Media, to appear in 2015. http:// dataintensive.net 2.  Jay Kreps: “Why local state is a fundamental primitive in stream processing.” 31 July 2014. http:// radar.oreilly.com/2014/07/why-local-state-is-a-fundamental-primitive-in-stream-processing.html 3.  Jay Kreps: “I ♥︎ Logs.” O'Reilly Media, September 2014. http://shop.oreilly.com/product/ 0636920034339.do 4.  Nathan Marz and James Warren: “Big Data: Principles and best practices of scalable realtime data systems.” Manning MEAP, to appear January 2015. http://manning.com/marz/ 5.  Jakob Homan: “Real time insights into LinkedIn's performance using Apache Samza.” 18 Aug 2014. http://engineering.linkedin.com/samza/real-time-insights-linkedins-performance-using-apache-samza 6.  Martin Kleppmann: “Moving faster with data streams: The rise of Samza at LinkedIn.” 14 July 2014. http://engineering.linkedin.com/stream-processing/moving-faster-data-streams-rise-samza-linkedin 7.  Praveen Neppalli Naga: “Real-time Analytics at Massive Scale with Pinot.” 29 Sept 2014. http:// engineering.linkedin.com/analytics/real-time-analytics-massive-scale-pinot 8.  Shirshanka Das, Chavdar Botev, Kapil Surlaker, et al.: “All Aboard the Databus!,” at ACM Symposium on Cloud Computing (SoCC), October 2012. http://www.socc2012.org/s18-das.pdf 9.  Apache Samza documentation. http://samza.incubator.apache.org 10. Alan Woodward and Martin Kleppmann: “Samza-Luwak Proof of Concept.” 10 November 2014. https://github.com/romseygeek/samza-luwak