Bay Area Apache Flink Meetup #2
Distributed Stream and Graph Processing
Community Update
August 2015
Henry Saputra
Committer and PMC Member
hsaputra@apache.org
@Kingwulf
Apache Flink is an open source platform for
scalable batch and stream data processing.
Apache Flink is …
2
• The core of Apache Flink is a
distributed streaming dataflow
engine.
• Executing dataflows in
parallel on clusters
• Providing a reliable
foundation for various
workloads
• DataSet and DataStream
programming abstractions are
the foundation for user programs
and higher layers
One engine for many use cases
3
Real time streaming
topologies
Machine Learning at scale
Graph Analysis
Long batch

pipelines
What happened? - 1
• New PMC: Maximilian Michels
• New Committer: Chesnay Schepler
• Discussions for a 0.9.1 release had started
• Apache Flink is becoming more popular:
– 1000+ Twitter followers
– 500+ GitHub stars
– Named as “open source Big Data project” to
watch by ZDNet.
– Flink Forward schedule with great speakers
announced
4
What happened? - 2
• Apache Flink on Wikipedia: https://
en.wikipedia.org/wiki/Apache_Flink
• New JobManager Dashboard
• Apache SAMOA 0.3.0-incubating with Flink
integration
• New “Features” page
• Contributors list (can you spot your name?)
https://cwiki.apache.org/confluence/display/
FLINK/List+of+contributors
5
New Job Manager Dashboard
6
New Website Redesign and
New Features page
7
New Architecture diagram in 0.10
documentation
8
More contents in the Wiki for
Internal Information
9
In master (0.10-SNAPSHOT) - 1
10
• Gelly Scala API
• More improvements and fixes for YARN
• Flink dropped Java 6 support
• Streaming connector for Elastic Search
• Sampling operation on DataSet API
• A lot of bug fixes:
– Streaming: APIs, general stability, kafka
connector
In master (0.10-SNAPSHOT) - 2
• Low watermarks / Event time
• New JM Dashboard
• Akka messages are now aware of leader
IDs (for HA)
• Zookeeper integration (for HA)
• Live accumulators (runtime only)
• Stability improvements
11
Articles and Mentions
• High-throughput, low-latency, and exactly-once stream
processing with Apache Flink [1]
• Introducing Gelly: Graph Processing with Apache Flink [2]
• Apache Flink and the case for stream processing [3]
• Crunching Parquet Files with Apache Flink [4]
• The morning paper: Asynchronous Distributed Snapshots for
Distributed Dataflows [5]
• Five open source Big Data projects to watch [6]
• Big Data Performance Engineering: Examples from Hadoop,
Pig, HBase, Flink and Spark [7]
12
[1] http://data-artisans.com/high-throughput-low-latency-and-exactly-once-stream-processing-with-apache-flink/
[2] http://flink.apache.org/news/2015/08/24/introducing-flink-gelly.html
[3] http://www.kdnuggets.com/2015/08/apache-flink-stream-processing.html
[4] https://medium.com/@istanbul_techie/crunching-parquet-files-with-apache-flink-200bec90d8a7
[5] http://blog.acolyer.org/2015/08/19/asynchronous-distributed-snapshots-for-distributed-dataflows/
[6] http://www.zdnet.com/article/five-open-source-big-data-projects-to-watch/
[7] http://www.bigsynapse.com/addressing-big-data-performance
New Meetups and Events
13
• Chicago: Flink Training @ Capital One
• Bay Area: Stream & Graph Processing @
MapR
13
GitHub stats
14
Upcoming
• Sept 15: Washington DC Area Apache
Flink Meetup
• Sept 17: StreamProcessing.be meetup
• Sept 28-30: Flink Talks at ApacheCon Big
Data Budapest
New Meetup groups:
• New York
• Boston
15
Flink Forward schedule published
16
• http://flink-forward.org/?post_type=day
• Talks by Google, Data Artisans, Huawei,
CapitalOne, Bouyges, Ericsson, Amadeus,
ResearchGate, RedHat, and many more.
50%
off for this meetup‘s guests
FlinkMeetupBayArea50

Bay Area Apache Flink Meetup Community Update August 2015

  • 1.
    Bay Area ApacheFlink Meetup #2 Distributed Stream and Graph Processing Community Update August 2015 Henry Saputra Committer and PMC Member hsaputra@apache.org @Kingwulf
  • 2.
    Apache Flink isan open source platform for scalable batch and stream data processing. Apache Flink is … 2 • The core of Apache Flink is a distributed streaming dataflow engine. • Executing dataflows in parallel on clusters • Providing a reliable foundation for various workloads • DataSet and DataStream programming abstractions are the foundation for user programs and higher layers
  • 3.
    One engine formany use cases 3 Real time streaming topologies Machine Learning at scale Graph Analysis Long batch
 pipelines
  • 4.
    What happened? -1 • New PMC: Maximilian Michels • New Committer: Chesnay Schepler • Discussions for a 0.9.1 release had started • Apache Flink is becoming more popular: – 1000+ Twitter followers – 500+ GitHub stars – Named as “open source Big Data project” to watch by ZDNet. – Flink Forward schedule with great speakers announced 4
  • 5.
    What happened? -2 • Apache Flink on Wikipedia: https:// en.wikipedia.org/wiki/Apache_Flink • New JobManager Dashboard • Apache SAMOA 0.3.0-incubating with Flink integration • New “Features” page • Contributors list (can you spot your name?) https://cwiki.apache.org/confluence/display/ FLINK/List+of+contributors 5
  • 6.
    New Job ManagerDashboard 6
  • 7.
    New Website Redesignand New Features page 7
  • 8.
    New Architecture diagramin 0.10 documentation 8
  • 9.
    More contents inthe Wiki for Internal Information 9
  • 10.
    In master (0.10-SNAPSHOT)- 1 10 • Gelly Scala API • More improvements and fixes for YARN • Flink dropped Java 6 support • Streaming connector for Elastic Search • Sampling operation on DataSet API • A lot of bug fixes: – Streaming: APIs, general stability, kafka connector
  • 11.
    In master (0.10-SNAPSHOT)- 2 • Low watermarks / Event time • New JM Dashboard • Akka messages are now aware of leader IDs (for HA) • Zookeeper integration (for HA) • Live accumulators (runtime only) • Stability improvements 11
  • 12.
    Articles and Mentions •High-throughput, low-latency, and exactly-once stream processing with Apache Flink [1] • Introducing Gelly: Graph Processing with Apache Flink [2] • Apache Flink and the case for stream processing [3] • Crunching Parquet Files with Apache Flink [4] • The morning paper: Asynchronous Distributed Snapshots for Distributed Dataflows [5] • Five open source Big Data projects to watch [6] • Big Data Performance Engineering: Examples from Hadoop, Pig, HBase, Flink and Spark [7] 12 [1] http://data-artisans.com/high-throughput-low-latency-and-exactly-once-stream-processing-with-apache-flink/ [2] http://flink.apache.org/news/2015/08/24/introducing-flink-gelly.html [3] http://www.kdnuggets.com/2015/08/apache-flink-stream-processing.html [4] https://medium.com/@istanbul_techie/crunching-parquet-files-with-apache-flink-200bec90d8a7 [5] http://blog.acolyer.org/2015/08/19/asynchronous-distributed-snapshots-for-distributed-dataflows/ [6] http://www.zdnet.com/article/five-open-source-big-data-projects-to-watch/ [7] http://www.bigsynapse.com/addressing-big-data-performance
  • 13.
    New Meetups andEvents 13 • Chicago: Flink Training @ Capital One • Bay Area: Stream & Graph Processing @ MapR 13
  • 14.
  • 15.
    Upcoming • Sept 15:Washington DC Area Apache Flink Meetup • Sept 17: StreamProcessing.be meetup • Sept 28-30: Flink Talks at ApacheCon Big Data Budapest New Meetup groups: • New York • Boston 15
  • 16.
    Flink Forward schedulepublished 16 • http://flink-forward.org/?post_type=day • Talks by Google, Data Artisans, Huawei, CapitalOne, Bouyges, Ericsson, Amadeus, ResearchGate, RedHat, and many more. 50% off for this meetup‘s guests FlinkMeetupBayArea50