ApacheCon is Coming 9-12 Sept. 2019 - Las Vegas The Apache Software Foundation
Apache 20th Anniversary Logo

Community-led development "The Apache Way"

Apache Support Logo

This was extracted (@ 2020-10-21 20:10) from a list of minutes which have been approved by the Board.
Please Note The Board typically approves the minutes of the previous meeting at the beginning of every Board meeting; therefore, the list below does not normally contain details from the minutes of the most recent Board meeting.

Meeting times vary, the exact schedule is available to ASF Members and Officers, search for "calendar" in the Foundation's private index page (svn:foundation/private-index.html).

Flink

16 Sep 2020 [Stephan Ewen / Shane]

## Description:
Apache Flink is a distributed data streaming system for batch and streaming
data analysis on top of a streaming dataflow engine. Flink's stack contains
functional batch and streaming analysis APIs in Java, Scala, Python, and
libraries for various use cases.

Flink interacts and integrates with several Apache projects in the broader
ecosystem of data storage and computing, such as Apache Beam, Calcite,
Hadoop, Mesos, Kafka, HBase, Cassandra, and various others.

## Issues:
 - There are currently no community issues that require board attention.

## Membership Data:
PMC changes since previous report:
- Yu Li joined the PMC on 2020-06-15
- Piotr Nowojski joined the PMC on 2020-06-15
- Dian Fu joined the PMC on 2020-08-27

New committers since previous report:
- Benchao Li was added as committer on 2020-06-08
- David Anderson was added as committer on 2020-08-08
- Yun Tang was added as committer on 2020-09-07
- Igal Shilman was added as committer on 2020-09-07

The newest PMC member is Dian Fu, joined on August 27th, 2020.
The newest committer is Igal Shilman, joined on September 7th, 2020.

There are currently 62 committers and 32 PMC members in this project.

## Releases
The following releases were made since the last board report:
 - 1.10.2 was released on 2020-08-25.
 - 1.11.0 was released on 2020-07-07.
 - 1.11.1 was released on 2020-07-22.

## Project Activity:
 - The Flink community is proud that Flink continues to be among the most
   active projects in the ASF. As per the FY 2020 Annual Report, Flink had
   the #1 most active dev@ mailing list (excluding builds, commits, and issue
   updates which go to separate lists) and the #2 most active user@ list,
   despite a growing second user mailing list (user-zh@).

 - The big 1.11.0 release went out in July, and a follow-up patch release
   since then as well. It included a lot of improvements, like a major update
   to Flink's checkpoint mechanism, support for change-data-capture streams,
   deeper Hadoop/Hive integration, better support for common Python libraries
   for Python users.

 - For the 1.12.0 release, the community has discussed the feature freeze to
   be end of October. The major release cycle is implicitly becoming a bit
   longer than it used to be, which seems a natural result due to the growth
   of the project, community, and issues going into releases.

 - The Stateful Functions project is heading for the 2.2 release, probably
   in the next two weeks, with several usability and stability improvements.

 - There will be a Virtual Flink Forward Community Conference, October 21-22.
   The main event is free for all.
   https://www.flink-forward.org/global-2020/conference-program

## Community Health:
After a very busy spring season, the summer season is slightly slower:

 - About 20 new major feature proposals (FLIPs = Flink Improvement Proposals)
 - user@f.a.o is roughly stable (2266 mails/quarter)
 - user-zh@f.a.o almost doubled (3359 mails/quarter)
 - dev@f.a.o is 25% slower (2340 mails/quarter)

Similarly, Jira activity is at 50% of last quarter;
 - 876 JIRA tickets created,
 - 782 JIRA tickets closed

This is probably explained by the very big 1.11 release to which the community
was gearing up in the last quarter, while this quarter didn't yet see the very
active phase leading to the next release (1.12 feature freeze is still two
months away).

17 Jun 2020 [Stephan Ewen / Sander]

## Description:
Apache Flink is a distributed data streaming system for batch and streaming
data analysis on top of a streaming dataflow engine. Flink's stack contains
functional batch and streaming analysis APIs in Java, Scala, Python, and
libraries for various use cases.

Flink interacts and integrates with several Apache projects in the broader
ecosystem of data storage and computing, such as Apache Beam, Calcite,
Hadoop, Mesos, Kafka, HBase, Cassandra, and various others.

## Issues:
 - There are currently no community issues that require board attention.

## Membership Data:
PMC changes since previous report:
- Dawid Wysakowicz joined PMC on 2020-03-23
- Zhijiang Wang joined the PMC on 2020-03-27
- Hequn Cheng joined the PMC on 2020-04-16

New committers since previous report:
- Konstantin Knauf was added as committer on 2020-04-01
- Seth Wiesman was added as committer on 2020-04-07
- Xintong Song was added as committer on 2020-06-01
- Benchao Li was added as committer on 2020-06-08

The newest PMC member is Hequn Cheng, joined on April 16th, 2020.
The newest committer is Benchao Li, joined on June 8th, 2020.

There are currently 59 committers and 29 PMC members in this project.

## Releases
The following releases were made since the last board report:
 - Apache Flink 1.9.3 was released on 2020-04-26.
 - Apache Flink 1.10.1 was released on 2020-05-12.
 - Stateful Functions (StateFun) 2.0 was released on 2020-04-07.
 - Stateful Functions (StateFun) 2.1 was released on 2020-06-09.
 - Flink-shaded 11.0 was released on 2020-05-29.

## Project Activity:
 - The community is currently finalizing the 1.11.0 release, the first
   release candidate has been published. Besides the common usability
   improvements and hardening, the release brings a major update to
   Flink's checkpoint mechanism, supports better handling of
   change-data-capture streams, integrates streaming data with Hadoop/Hive,
   adds better support for common libraries for Python users, and more.

 - The contributed Stateful Functions project has now undergone two releases
   as part of Apache Flink, meaning the integration has been successfully
   completed.

 - The Flink Forward San Francisco conference took place as an open virtual
   event due to the COVID-19 situation. While this worked well for the part
   of knowledge and experience sharing, the feedback from many committers
   was that they prefer an in-person event once it becomes feasible again.

 - For the same reason as above, the next Flink Forward conference (taking
   place in late October) will be again a virtual event, open and free to
   everyone.

 - There is a request to use the Apache Flink brand for a fan page. The PMC
   is handling the discussion. The main point is how to strike a good trade-off
   between making it clear that this is not an official website of the
   Apache Flink project, and making it recognizable at first glance that this
   website is about Flink.

 - The Flink community migrated most of the CI infrastructure to a sponsored
   Azure Pipelines account. That reduced Flink's usage of ASF Travis resources
   significantly. The remaining ASF Travis CI use will probably be phased out
   incrementally along with older releases that still rely on the Travis setup.

 - The community published its first CVE based on a third-party report and
   fix. We found the process to be well documented.

 - There is an ongoing discussion in the community about adding more
   localized contents, triggered by the contribution of a Japanese translation
   of (parts of) the website. At this point, the conclusion was that we
   need to first invest into better tooling before being able to handle
   more versions of the documentation (website being a bit easier, because
   it is less frequently updated).

## Community Health:
The Flink community is proud that Flink continues to be among the most active
projects in the ASF.
 - user@f.a.o (2452 mails/quarter)
 - user-zh@f.a.o (1930 mails/quarter)
 - dev@f.a.o (3914 mails/quarter)

JIRA continues to be very active as well, 1622 JIRA tickets created,
1351 JIRA tickets closed in the last 3 months.

18 Mar 2020 [Stephan Ewen / Danny]

## Description:
Apache Flink is a distributed data streaming system for batch and streaming
data analysis on top of a streaming dataflow engine. Flink's stack contains
functional batch and streaming analysis APIs in Java, Scala, Python, and
libraries for various use cases.

Flink interacts and integrates with several Apache projects in the broader
ecosystem of data storage and computing, such as Apache Beam, Calcite,
Hadoop, Mesos, Kafka, HBase, Cassandra, and various others.

## Issues:
  - There are no community issues that require board attention.

## Membership Data:

No new PMC members were added since the last report.

Committers added since the last board report:
- Yu Li was added as committer on January 7th, 2020
- Dian Fu was added as committer on January 17th, 2020
- Jingsong Lee was added as committer on February 2nd, 2020

The newest PMC member is Jark Wu, joined on November 8th, 2019.
The newest committer is Jingsong Lee, joined on February 2nd, 2020.

There are currently 55 committers and 26 PMC members in this project.

## Releases
The following releases were made since the last board report:
 - Flink 1.9.2 was released on January 31st, 2020.
 - Flink 1.10.0 was released on February 11th, 2020.
 - Flink Shaded 10.0 (vendored libraries) was released on February 17th, 2020

## Project Activity:
  - The community has released Flink 1.10, a big release with many features,
    notably easier memory configuration, Python functions, an active Kubernetes
    integration, integration with Hive and fulls TPC-DS coverage in SQL.

  - The ASF has registered Flink as a trademark for the EU and China regions
    now. Flink was already registered as a trademark in the US.

  - The contributed Stateful Functions project is on its way to its first
    release under its new home in the Flink community. Code and documentation
    has been successfully migrated, discussions are ongoing about updating
    the presentation of the website to better reflect the Flink project's
    evolution.

  - The Flink Forward community conference has been canceled due to
    the Covid-19 outbreak, and the organizers are exploring options to
    have an online event instead.

  - Discussions about the 1.11 release have been started by the designated
    release managers.

  - The community is exploring a "developer blog" to better exchange
    information about internals of the system and project between contributors,
    that might not be very relevant or interesting to most users.


## Community Health:
The mailing list activity remains stable on a high level.
  - user@f.a.o (2027 mails/quarter)
  - user-zh@f.a.o (706 mails/quarter)
  - dev@f.a.o (2997 mails/quarter)

JIRA continues to be very active as well, 1315 JIRA tickets created,
843 JIRA tickets closed in the last 3 months.

18 Dec 2019 [Stephan Ewen / Ted]

## Description:
Apache Flink is a distributed data streaming system for batch and streaming
data analysis on top of a streaming dataflow engine. Flink's stack contains
functional batch and streaming analysis APIs in Java, Scala, Python, and
libraries for various use cases.

Flink interacts and integrates with several Apache projects in the broader
ecosystem of data storage and computing, such as Apache Beam, Calcite, Hadoop,
Mesos, Kafka, HBase, Cassandra, and various others.

## Issues:
 - There are no community issues that require board attention.

## Membership Data:
PMC members added since the last board report:
 - Jark Wu joined the PMC on November 8th, 2019
 - Becket Qin joined the PMC on October 21st, 2019

Committers added since the last board report:
 - Zhu Zhu was added as committer on December 17th, 2019

The newest PMC member is Jark Wu, joined on November 8th, 2019. The newest
committer is Zhu Zhu, joined on December 17th, 2019.

Flink currently has 52 committers and 26 PMC members.

## Releases
The following releases were made since the last board report:

 - 1.8.3 was released on December 12th, 2019.
 - 1.9.1 was released on October 18th, 2019.
 - shaded-9.0 was released on November 23rd, 2019.

## Project Activity:
 - The community has frozen the feature set for the 1.10 release and is now
   in the testing and validation phase. The release adds various enhancements
   to the runtime (easier memory configuration and threading model) and many
   enhancements to the "batch on streaming" capabilities of Flink, including
   the Table API/SQL and integrations with the SQL/Hive ecosystem.

 - There is an ongoing discussion how to better handle the use of the Apache
   Flink name in China, where there is a large community. The ASF does not
   have the trademark for Flink in China. The suggestion from Apache
   trademarks is that the PMC does outreach and education for now, and not
   push for a trademark registration process at the moment.

 - The PMC voted to start an WeChat account for Flink that is managed by the
   PMC. Similar to the @ApacheFlink Twitter account, this channel would be
   used for various types of Flink news and to publicize interesting
   mentionings of Flink from users and contributors.

 - The Flink community voted to accept the "Stateful Functions" project as
   contribution. https://statefun.io/ Stateful Functions is a library built
   on Apache Flink to support new use cases that are not easily mappable to
   the existing APIs, but more of a match distributed actor frameworks. The
   library was already open source and Apache 2 licensed.

 - The community is exploring to migrate the CI service fully away from
   Travis CI. Currently Pull Requests are tested through a separately-paid
   Travis subscription to take load off the ASF-paid account/budget, which
   is only used to test the master branch.
   The goal is to have everything in Azure Pipelines, which offers better
   tools as well as the opportunity to hook in external build servers that
   Alibaba is offering to sponsor. We hope that this both simplifies Flink's
   build setup, as well as reduces feedback times on Pull Requests
   significantly.

 - The next Flink Forward community conference will be in San Francisco on
   March 23rd-25th. There is a public call for presentations:
   https://www.flink-forward.org/sf-2020/call-for-presentations Apache
   committers can contact the organizers for a free pass.

## Community Health:
After a big increase in activity in the previous quarter, the mailing list
activity seems to be stable now.

 - user@f.a.o (1730 mails/quarter)
 - user-zh@f.a.o (570 mails/quarter)
 - dev@f.a.o (3057 mails/quarter)

JIRA continues to be active as well, 1096 JIRA tickets created, 833 JIRA
tickets resolved in the last 3 months.

18 Sep 2019 [Stephan Ewen / Shane]

## Description:

Apache Flink is a distributed data streaming system for batch and streaming
data analysis on top of a streaming dataflow engine. Flink's stack contains
functional batch and streaming analysis APIs in Java, Scala, Python, and
libraries for various use cases.

Flink interacts and integrates with several Apache projects in the broader
ecosystem of data storage and computing, such as Apache Beam, Calcite,
Hadoop, Mesos, Kafka, HBase, Cassandra, and various others.

## Issues:

 - There are no issues that require board attention.

## Membership Data:

PMC members added since the last board report:
 - Jincheng Sun joined the PMC on June 24th, 2019
 - Kurt Young joined the PMC on July 23rd, 2019
 - Kostas Kloudas joined the PMC on August 31st, 2019

Committers added since the last board report:
 - Rong Rong was added as committer on July 15th, 2019
 - Becket Qin was added as committer on July 18th, 2019
 - Zhijiang Wang was added as committer on July 22nd, 2019
 - Hequn Cheng was added as committer on August 5th, 2019
 - Andrey Zagrebin was added as committer on August 17th, 2019
 - Zili Chen was added as committer on September 12th, 2019

The newest PMC member is Kostas Kloudas, joined on August 31st, 2019.
The newest committer is Zili Chen, joined on September 12th, 2019.

Flink currently has 51 committers and 24 PMC members.

## Releases

The following releases were made since the last board report:

 - 1.8.1 was released on July 7th, 2019.
 - 1.8.2 was released on September 11th, 2019
 - 1.9.0 was released on August 22nd, 2019.
 - shaded-8.0 was released on August 8th, 2019.

## Project Activity:

 - The community is in a strong growth phase. The past quarter saw the
   addition of six new committers and three new PMC members to help support
   the community growth.

 - The Flink community passed its first set of bylaws to more formally define
   how decisions are made for various types of contributions. The community
   felt that with growing size of active committer base and with the
   different backgrounds and experiences of the committers, bylaws will be
   helpful to keep the project organized.

 - Parts of the bylaws discussion resulted also in reinforcing the process
   for new features (FLIPs - Flink Improvement Proposals), which resulted in
   a big number of new proposals and high activity on their discussion
   threads.

 - Also for smaller contributions, the community is pushing more for early
   discussions and checking whether there is capacity among committer to
   shepherd proposed changes.

 - The community is periodically summing up the ongoing development threads
   on the roadmap page: https://flink.apache.org/roadmap.html

 - The community started working on the Flink 1.10 release. The release will
   continue many of the threads started for 1.9, like the Table API, SQL
   engine, integration with Hive, support for Python. Other threads include,
   for example, a new streaming state backend, sources, as well as making
   configuration and resource management simpler.

 - Flink Forward Berlin will be on October 7-9. The program is online:
   https://europe-2019.flink-forward.org/speakers Apache committers can
   contact the organizers for a free pass.

## Community Health:

Mailing lists continue to be very active:
 - user@f.a.o (1767 mails/quarter)
 - user-zh@f.a.o (714 mails/quarter)
 - dev@f.a.o (3543 mails/quarter)

The dev mailing list saw a big increase in activity due to feature and design
discussions for the 1.10 release and the reinforced community processes.

The user-zh mailing list seems to be building momentum as well and more
than doubled in the last quarter.

JIRA continues to be active as well, 1212 JIRA tickets created, 910 JIRA
tickets resolved in the last 3 months.

19 Jun 2019 [Stephan Ewen / Danny]

DESCRIPTION

Apache Flink is a distributed data streaming system for batch and streaming
data analysis on top of a streaming dataflow engine. Flink's stack contains
functional batch and streaming analysis APIs in Java and Scala and libraries
for various use cases.

Flink interacts and integrates with several Apache projects in the broader
ecosystem of data storage and computing, such as Apache Beam, Calcite, Hadoop,
Mesos, Kafka, HBase, Cassandra, and various others.

ISSUES

  - There are no issues that require board attention.

STATUS AND ACTIVITY

  - The work on improving community aspects is still ongoing. There is work in
    reorganizing the contribution process to encourage more early discussions,
    rather than immediately opening pull requests.

  - There are now more regular discussions on the private mailing list about
    committer and PMC candidates now. The project is adding PMC members and
    committers more regularly now. Actively reaching out and mentoring new
    candidate needs more follow-up work by the PMC.

  - The integration of the contributed Blink fork is ongoing. A good part of
    it will be included in the next Flink release (1.9).

  - The community working on the Flink 1.9 release. The release has among
    other things a complete overhaul of the Table API, a new SQL/Table query
    processor based on the Blink contribution, a new scheduler architecture
    and failover mechanics for batch/bounded programs. The release also
    introduces a first version of Python Table API.

  - Flink Forward San Francisco took place on April 1-2 with many Flink
    committers and contributors attending. The call for participation for
    Flink Forward Berlin just closed, the organizers plan to offer free passes
    again for Flink committers.

COMMUNITY

PMC members added since the last board report:
 - Shaoxuan Wang joined the PMC on May 15th, 2019

Committers added since the last board report;
 - Bowen Li was added as a committer on April 22nd, 2019
 - Xiaowei Jiang was added as a committer on April 1st, 2019.

The newest PMC member is  Shaoxuan Wang, joined on May 15th, 2019. The newest
committer is  Bowen Li, joined on  April 22nd, 2019.

Flink currently has 45 committers and 21 PMC members.

RELEASES

The following releases were made since the last board report:

 - 1.8.0 was released on April 9th, 2019
 - shaded-7.0 was released on May 30th, 2019

ACTIVITY ON JIRA / MAILING LISTS

Mailing lists continue to be very active:
  - user@f.a.o (1634 mails/quarter)
  - dev@f.a.o (1672 mails/quarter)

JIRA continues to be active as well, 918 JIRA tickets created, 624 JIRA
tickets resolved in the last 3 months.

20 Mar 2019 [Stephan Ewen / Brett]

DESCRIPTION

Apache Flink is a distributed data streaming system for batch and streaming
data analysis on top of a streaming dataflow engine. Flink's stack contains
functional batch and streaming analysis APIs in Java and Scala and libraries
for various use cases.

Flink interacts and integrates with several Apache projects in the broader
ecosystem of data storage and computing, such as Apache Beam, Calcite,
Hadoop, Mesos, Kafka, HBase, Cassandra, and various others.

ISSUES

  - There are no issues that require board attention.

STATUS AND ACTIVITY

  - The Flink community is trying to improve many community aspects in
    the last months, resulting in a lot of discussions and efforts.

  - Inspired by the effort from the Apache Beam project, the Flink PMC
    is looking to have a monthly discussion on promising candidates
    for committers and PMC and actively reach out. There are already
    healthy discussions about PMC and committer growth.

  - The community added a Chinese language user@flink mailing list for
    user-to-user questions and support. Since its creation six weeks ago,
    the list has already gathered 300 subscribers.
    Design discussions and decisions still happen on the dev mailing list.

  - The trademark issue with the https://flink-china.org website (mainly a
    translation of the official Flink website) was resolved by convincing
    the maintainers to contribute the contents to the Apache website.
    The Flink website is now dual-language (Chinese version under
    https://flink.apache.org/zh/).
    The https://flink-china.org is currently online, with a disclaimer that
    it is being contributed and to the official Flink website and will then
    be shut down.

  - Alibaba has contributed the code of their internal Flink fork 'Blink' back
    to the open source project. The code is available in a separate branch,
    the community is working on merging selected parts into the mainline.

  - The community published a roadmap for the project, which sums up the
    active Flink Improvement Proposals (FLIPs) and JIRA issues to give
    users and contributors an easily digestible overview of all the ongoing
    efforts.

  - There are several discussion threads on how to improve the experience of
    contributing to Flink and keeping up with a large number of contributions.
    Current efforts include a better organization of JIRA components,
    a more transparent review process, and some automation through a pull
    request bot. In addition, we are thinking about component-specific guides
    and test frameworks, for example for connectors.

  - The community is in the process of releasing Flink 1.8. The release
    contains some new user-facing features plus a lot of internal cleanup
    and refactoring, fixing some long term issues, and preparing the work
    on a deeper integration of batch and stream processing in the runtime.

  - Program for Flink Forward San Francisco is online:
    https://sf-2019.flink-forward.org/conference-program
    Flink committers can get a free pass, and there are a discount codes for
    other Apache committers.

COMMUNITY

PMC members added since the last board report:
 - Thomas Weise joined the PMC on February 12th, 2019

No new committers were added since the last board report.

The newest PMC member is Thomas Weise, joined on February 12th, 2019.
The newest committer is Jamie Grier, joined on December 17th, 2018.

Flink currently has 43 committers and 20 PMC members.

RELEASES

The following releases were made since the last board report:

 - 1.6.3 was released on December 21st, 2018
 - 1.6.4 was released on February 22nd, 2019
 - 1.7.1 was released on December 20th, 2018
 - 1.7.2 was released on February 14th, 2019
 - shaded-6.0 was released on February 12th, 2019

ACTIVITY ON JIRA / MAILING LISTS

Mailing lists continue to be very active:
  - user@f.a.o (1737 mails/quarter)
  - dev@f.a.o (1963 mails/quarter)

JIRA continues to be active as well, 767 JIRA tickets created, 836 JIRA
tickets resolved in the last 3 months.

19 Dec 2018 [Stephan Ewen / Brett]

DESCRIPTION

Apache Flink is a distributed data streaming system for batch and streaming
data analysis on top of a streaming dataflow engine. Flink's stack contains
functional batch and streaming analysis APIs in Java and Scala and libraries
for various use cases.

Flink interacts and integrates with several Apache projects in the broader
ecosystem of data storage and computing, such as Apache Beam, Calcite,
Hadoop, Mesos, Kafka, HBase, Cassandra, and various others.

ISSUES

  - There are no issues that require board attention.

STATUS AND ACTIVITY

  - The Flink community has released the 1.7 release, with features such
    as support for Scala 2.12, enhancements to Streaming SQL, Complex Event
    processing, AWS S3, connectors, and recovery.

  - The project has entered a phase of many discussions about breaking
    changes and consolidation of batch and streaming APIs to fix some
    shortcomings and technical complexity, due to the way the project evolved
    organically over time.

  - In the area of Streaming SQL, there is a fruitful collaboration with the
    Apache Calcite project, which resulted for example in new streaming joins
    and an integration of complex event processing and SQL.

  - The program for the Flink Forward Beijing conference is available. Aside
    from tech talks, there will be a keynote by Craig Russell about the ASF
    and how to contribute to Apache projects.

COMMUNITY

No new PMC members were added since the last report.
The newest PMC member is Chesnay Schepler, joined on July 26th, 2017

Committers added since the last board report:

 - Thomas Weise was added as a committer on November 21st, 2018
 - Jamie Grier was added as a committer on December 13th, 2018

Flink currently has 43 committers and 19 PMC members.

RELEASES

The following releases were made since the last board report:

 - 1.5.4 was released on September 20th, 2018
 - 1.5.5 was released on October 29th, 2018
 - 1.6.1 was released on September 20th, 2018
 - 1.6.2 was released on October 29th, 2018
 - 1.7.0 was released on November 30th, 2018

ACTIVITY ON JIRA / MAILING LISTS

Mailing lists continue to be very active:
  - user@f.a.o (2239 mails/quarter)
  - dev@f.a.o (1825 mails/quarter)

JIRA continues to be active as well, 807 JIRA tickets created, 569 JIRA
tickets resolved in the last 3 months.

19 Sep 2018 [Stephan Ewen / Brett]

DESCRIPTION

Apache Flink is a distributed data streaming system for batch and streaming
data analysis on top of a streaming dataflow engine. Flink's stack contains
functional batch and streaming analysis APIs in Java and Scala and libraries
for various use cases.

Flink interacts and integrates with several Apache projects in the broader
ecosystem of data storage and computing, such as Apache Beam,
Hadoop, Mesos, Kafka, HBase, Cassandra, and various others.

ISSUES

  - There are no issues that require board attention.

STATUS AND ACTIVITY

  - The Flink community has released the 1.6 release, with enhancements
    to Streaming SQL, connectors, container-based deployment, and more.

  - The investment in more automated release testing seems to pay off. The
    community is releasing bugfix releases faster and more frequently.

  - The Flink Forward Berlin conference took place earlier in September April
    bringing together many community members. Talks were a mix of use cases,
    tech deep dive, ecosystem, and (for the first time) stream processing
    research topics.

  - Two more Flink Forward conferences have been announced: In Beijing on
    December 20-21, 2018, and in San Francisco on April 1-2, 2019.

COMMUNITY

No new PMC members were added since the last report.
The newest PMC member is Chesnay Schepler, joined on July 26th, 2017

Committers added since the last board report:
  - Sihua Zhou was added as a committer on June 23rd, 2018
  - Piotr Nowojski was added as a committer on June 27th, 2018
  - Gary Yao was added as a committer on September 9th, 2018

Flink currently has 41 committers and 19 PMC members.

RELEASES

The following releases were made since the last board report:

 - 1.5.1 was released on July 12th, 2018
 - 1.5.2 was released on July 31st, 2018
 - 1.5.3 was released on August 21st, 2018
 - 1.6.0 was released on August 9th, 2018

ACTIVITY ON JIRA / MAILING LISTS

Mailing lists continue to be very active and steadily grow in traffic:
  - user@f.a.o (2376 mails/quarter)
  - dev@f.a.o (1413 mails/quarter)

JIRA continues to be active as well, 740 JIRA tickets created, 565 JIRA
tickets resolved in the last 3 months .

20 Jun 2018 [Stephan Ewen / Isabel]

DESCRIPTION

Apache Flink is a distributed data streaming system for batch and streaming
data analysis on top of a streaming dataflow engine. Flink's stack contains
functional batch and streaming analysis APIs in Java and Scala and libraries
for various use cases.

Flink interacts and integrates with several Apache projects in the broader
ecosystem of data storage and computing, such as Apache Beam,
Hadoop, Mesos, Kafka, HBase, Cassandra, and various others.

ISSUES

 - There are no issues that require board attention.

STATUS AND ACTIVITY

 - The Flink community has released the 1.5 release, with major changes to the
   deployment model, network stack, recovery, and more.

 - Community has invested a lot of time into automating release testing, to
   be able to switch to a more strict time-based release model.

 - Work towards the next version (1.6) has started, a discussion about parts of
   the feature set is still ongoing on the mailing lists.

 - The Flink Forward San Francisco conference took place in April,
   with a mix of committers, contributors, users, and explorers.
   The call for presentations for the next Flink Forward conference
   in Berlin (September 3rd-5th) is open.

COMMUNITY

No new PMC members were added since the last report.
The newest PMC member is Chesnay Schepler, joined on July 26th, 2017

Committers added since the last board report:
 - Nico Kruber was added as a committer on May 10th, 2018
 - Xingcan Cui was added as a committer on May 10th, 2018
 - Shuyi Chen was added as a committer on June 5th, 2018

Flink currently has 38 committers and 19 PMC members.

RELEASES

The following releases were made since the last board report:

 - 1.5.0 was released on May 25th, 2018
 - shaded-4.0 (auxiliary project) was released on June 6th, 2018

ACTIVITY ON JIRA / MAILING LISTS

The Flink community is proud to continues to be in the top 10
active mailing lists in Apache:
 - user@f.a.o (1847 mails/quarter)
 - dev@f.a.o (1285 mails/quarter)

JIRA continues to be active as well, 653 JIRA tickets created, 589 JIRA
tickets resolved in the last 3 months .

21 Mar 2018 [Stephan Ewen / Phil]

DESCRIPTION

Apache Flink is a distributed data streaming system for batch and streaming
data analysis on top of a streaming dataflow engine. Flink's stack contains
functional batch and streaming analysis APIs in Java and Scala and libraries
for various use cases.

Flink interacts and integrates with several Apache projects in the broader
ecosystem of data storage and computing, such as Apache Beam,
Hadoop, Mesos, Kafka, HBase, Cassandra, and various others.

ISSUES

 - There are no issues that require board attention.

STATUS AND ACTIVITY

 - The Flink community is in the final phase of the 1.5 release, with the
   release branch being forked and final tests and bug fixes happening.

 - The 1.5 release marks a major rework of Flink, to make its design more
   natural for idiomatic use with containers (specifically Kubernetes) and more
   dynamic resource environments, but at the same time retain the current
   integration with Yarn, Mesos, and standalone setups.
   Aside from that, the release improves latency, checkpoints, recovery time,
   and extends streaming SQL and the DataStream API, among other things.

 - The Flink Forward San Francisco conference will happen in three weeks.
   Flink committers were offered a free pass.
   The program can be found under https://sf-2018.flink-forward.org/conference/

 - Concerning the note in the previous board report, the Huawei
   CloudStream website now uses Apache Flink, rather than only
   Flink (accessed March 15th)
   https://www.huaweicloud.com/en-us/product/cs.html

COMMUNITY

No new PMC members were added since the last report.
The newest PMC member is Chesnay Schepler, joined on July 26th, 2017

Committers added since the last board report:
 - Eron Wright was added as a committer on January 17th, 2018

Flink currently has 35 committers and 19 PMC members.

RELEASES

The following releases were made since the last board report:

 - 1.4.1 was released on February 15th, 2018
 - 1.4.2 was released on March 8th, 2018
 - 1.3.3 was released on March 15th, 2018
 - shaded-3.0 was released on February 28th, 2018

ACTIVITY ON JIRA / MAILING LISTS

The Flink community was very proud to be in the top 10 active mailing lists in Apache :-)

user@f.a.o (1817 mails/quarter) and dev@f.a.o (1154 mails/quarter)

As the community pushes towards the 1.5 release, JIRA continues
to be very active, with 688 JIRA tickets created and 672 JIRA tickets
resolved over last 3 months.

20 Dec 2017 [Stephan Ewen / Phil]

DESCRIPTION

Apache Flink is a distributed data streaming system for batch and streaming
data analysis on top of a streaming dataflow engine. Flink's stack contains
functional batch and streaming analysis APIs in Java and Scala and libraries
for various use cases.

Flink interacts and integrates with several Apache projects in the broader
ecosystem of data storage and computing, such as Apache Beam,
Hadoop, Mesos, Kafka, HBase, Cassandra, and various others.

ISSUES

  - There are no issues that require board attention.

STATUS AND ACTIVITY

  - The Flink community just released the 1.4 release and is
    now working towards the 1.5 release. The relationship between
    the releases and a list of major features has been has been
    discussed and is shared on the Apache Flink blog:
    http://flink.apache.org/news/2017/11/22/release-1.4-and-1.5-timeline.html

  - The call for papers for the Flink Forward conference is open.
    The program committee consists of stream processing experts from
    both the Flink community and other communities.

  - Flink had a pretty good conference presence in the previous
    quarter, for example at multiple Stratas, Hadoop/Dataworks Summit,
    MesosCon, QCon, Scale-by-the-Bay, and Flink Forward Berlin.

  - We reached out to Huawei to update the use of "Flink" on their stream
    product website to "Apache Flink". The original Chinese webpage was fixed,
    but has been taken down and replaced by other web pages, including an
    English one that uses the original text saying "Flink", not "Apache Flink".
    We are in touch.


COMMUNITY

No new PMC members were added since the last report.
The newest PMC member is Chesnay Schepler, joined on July 26th, 2017

Committers added since the last board report:
  - Haohui Mai was added as a committer on November 1st, 2017

Flink currently has 34 committers and 19 PMC members.


RELEASES

The following releases were made since the last board report:

 - 1.4.0 was released on December 12th, 2017
 - shaded-2.0 was released on October 30th, 2017


ACTIVITY ON JIRA / MAILING LISTS

Both user@f.a.o and dev@f.a.o lists are very active,
(1818 resp. 977 mails per quarter), with mails getting
answered by a mix of committers and non-committers.

JIRA continues to be very active, with 636 JIRA tickets created
and 607 JIRA tickets resolved over last 3 months.

20 Sep 2017 [Stephan Ewen / Rich]

DESCRIPTION

Apache Flink is a distributed data streaming system for batch and streaming
data analysis on top of a streaming dataflow engine. Flink's stack contains
functional batch and streaming analysis APIs in Java and Scala and libraries
for various use cases.

Flink interacts and integrates with several Apache projects in the broader
ecosystem of data storage and computing, such as Apache Beam,
Hadoop, Mesos, Kafka, HBase, Cassandra, and various others.

ISSUES

 - There are no issues that require board attention.

STATUS AND ACTIVITY

 - The Flink community is heading towards the 1.4 release. Big
   threads in the current development are for example a major
   improvement to the deployment interoperability and elasticity,
   connectors, dependency management, Stream SQL.

 - The Flink community is reorganizing the project repository and
   dependencies, to make the life of users easier with respect to
   dependency conflicts. We created the "flink-shaded" repository
   in which we create shaded versions of several dependencies,
   relocated to a different namespace. That repository contains at
   this point no code, only build files.

 - The Flink Forward Berlin conference took place on September 11-13
   with about 350 participants, representing a mixture of committers,
   contributors, production users, exploring users, stream processing
   enthusiasts, and researchers from academia.

 - Data Artisans announced that they plan to organise two more
   Flink Forward conferences next year, similar to this year:
   One in San Francisco in April, one in Berlin in September.

 - Since the last board report, two companies publicly annouced
   hosted managed Flink services: Eventador (http://eventador.io/)
   and Huawei (http://www.hwclouds.com/product/stream.html).
   We are in touch with the later one to update their mentioning
   of "Flink" to "Apache Flink".

COMMUNITY

PMC members added since last report:
 - Tzu-Li Tai was added as a PMC member on July 6th, 2017
 - Chesnay Schepler was added as a PMC member on July 26th, 2017

Committers added since the last board report:
 - Jincheng Sun was added as a committer on July 10th, 2017

Flink currently has 33 committers and 19 PMC members.

RELEASES

The following releases were made since the last board report:
 - 1.3.1 was released on June 22nd, 2017
 - 1.3.2 was released on August 5th, 2017
 - shaded-1.0 was released on July 27th, 2017

ACTIVITY ON JIRA / MAILING LISTS

Both user@f.a.o and dev@f.a.o lists are very active, with mails getting
answered by a mix of committers and non-committers.

JIRA continues to be very active
(699 JIRA tickets created, 538 JIRA tickets resolved over last 3 months)

21 Jun 2017 [Stephan Ewen / Chris]

DESCRIPTION

Apache Flink is a distributed data streaming system for batch and streaming
data analysis on top of a streaming dataflow engine. Flink's stack contains
functional batch and streaming analysis APIs in Java and Scala and libraries
for various use cases.

Flink interacts and integrates with several Apache projects in the broader
ecosystem of data storage and computing, such as Apache Beam,
Hadoop, Mesos, Kafka, HBase, Cassandra, and various others.

ISSUES

 - There are no issues that require board attention.

STATUS AND ACTIVITY

 - Flink is in a healthy phase of growing the committer base to handle the
   increased activity in various components. Since the beginning
   of the year, eight new committers were added.

 - The first time-based release happened with 1.3.0, and the feedback
   from users and committers was positive.

 - The discussion about modularization and possible changes to the build and
   testing setups continues, to address the very long build/test times.
   No conclusion has been reached so far. A significant build-time reduction
   would require big changes to the build setups (like organizing Flink into
   various sub-repositories) and there are concerns about the added complexity
   for users and contributors that such a setup brings.

 - There will be another instance of the Flink Forward conference in Berlin
   in September this year: http://berlin.flink-forward.org
   The program committee consists of a mix of Flink community members and
   stream processing experts from other stream and event-processing communities.
   The call for participation is currently happening.

 - The Flink Forward San Francisco conference happened in April, with
   Apache Flink training, user talks, committer talks, and a "how to
   participate in the community" tutorial.
   In addition, there were presentations about Flink at Apache Con,
   Kafka Summit, Dataworks Summit, Strata, Mesos Con, J on the Beach,
   and several other conferences.

 - Voting on bugfix release 1.3.1 is in progress at the time of writing
   this report.

COMMUNITY

PMC members added since last report:
 - Greg Hogan was added as a PMC member on March 13th, 2017.

Committers added since the last board report:
 - Theodore Vasiloudis was added as a committer on March 21st, 2017
 - Xiaogang Shi was added as a committer on May 21st, 2017
 - Dawid Wysakowicz was added as a committer on June 14th, 2017
 - Shaoxuan Wang was voted in as a committer on June 15th, 2017

Flink currently has 32 committers and 17 PMC members.

RELEASES

The following releases were made since the last board report:

 - 1.1.5 was released on March, 22nd, 2017
 - 1.2.1 was released on April 25th, 2017
 - 1.3.0 was released on May 31st, 2017

ACTIVITY ON JIRA / MAILING LISTS

Both user@f.a.o and dev@f.a.o lists are very active, with mails getting
answered by a mix of committers and non-committers.

JIRA continues to be very active
(824 JIRA tickets created, 622 JIRA tickets over last 3 months)

15 Mar 2017 [Stephan Ewen / Isabel]

DESCRIPTION

Apache Flink is a distributed data streaming system for batch and streaming
data analysis on top of a streaming dataflow engine. Flink's stack contains
functional batch and streaming analysis APIs in Java and Scala and libraries
for various use cases.

Flink interacts and integrates with several Apache projects in the broader
ecosystem of data storage and computing, such as Apache Beam (incubating),
Hadoop, Kafka, HBase, Cassandra, and others.

ISSUES

 - There are no issues that require board attention.

STATUS AND ACTIVITY

 - The project continues to have quite a high activity, with many
   contributions both for complex long-term features, and small usability-
   and bug fixes.  We are growing the committer base in order to keep up
   with the contributions.

 - While additional committers are helping in many of the user-facing
   components, there are still some more low-level components where
   expertise naturally takes a bit longer to spread. The committers keep
   working with more contributors on those parts to help with that.

 - The PMC is currently in progress of adding another member. The vote has
   passed, currently waiting for the board notice period before extending
   the invitation.

 - After the previous release, the community discussed and decided to
   switch to a strictly time-based release policy. That makes it
   more predictable for both users and contributors. The initial
   interval was decided to be 4 months, which would mark the next
   release at mid May.

 - The community has started to gather issues for the bugfix releases
   1.2.1 and probably 1.1.5

 - There has been a long discussion (not fully concluded) how to make it
   easier to working on Flink libraries (Graph Processing, Complex Event
   Processing, Machine Learning). Some issues are teh large code base and
   the increasingly long build times. Other discussions are how to foster
   activity in those specific components.

 - The Program for the upcoming Flink Forward conference in San Francisco
   has been announced. There will be talks by users and committers about
   use cases, technology, community participation, and ecosystem
   integration.

 - Aside from the talks at Flink Forward, there will be presentations about
   Flink at Strata, Dataworks Summit, Kafka Summit, and other conferences.

 - The results from a Flink survey by data Artisans have been shared in a
   two part blog post. The link to the results was was shared over the
   user/dev mailing lists.


COMMUNITY

We are in the process of adding a new PMC member.
The latest PMC addition was July 2nd, 2015 (Maximilian Michels)

Committers added since the last board report:
 - Kostas Kloudas was added as a committer on February, 5th, 2017
 - Jark Wu was added as a committer on February, 9th, 2017
 - Stefan Richter was added as a committer on February 9th 2017
 - Kurt Young was added as a committer on February 21st, 2017

Flink currently has 28 committers and 16 PMC members.

RELEASES

The following releases were made since the last board report:

 - 1.1.4 was released on December, 21st, 2016
 - 1.2.0 was released on February 2nd, 2017

ACTIVITY ON JIRA / MAILING LISTS

Both user@f.a.o and dev@f.a.o lists are very active, with mails getting
answered by a mix of committers and non-committers.

JIRA continues to be very active
(713 JIRA tickets created, 630 JIRA tickets over last 3 months)

21 Dec 2016 [Stephan Ewen / Brett]

DESCRIPTION

Apache Flink is a distributed data streaming system for batch and streaming
data analysis on top of a streaming dataflow engine. Flink's stack contains
functional batch and streaming analysis APIs in Java and Scala and libraries
for various use cases.

Flink interacts and integrates with several Apache projects in the broader
ecosystem of data storage and computing, such as Apache Beam (incubating),
Hadoop, Kafka, HBase, Cassandra, Bahir, and others.

ISSUES

  - There are no issues that require board attention.

STATUS AND ACTIVITY

  - The Flink community is heading towards the 1.2 release, which is the
    biggest release so far. The community decided to add backwards
    compatibility of "application state" from this release on (in addition
    to API compatibility since version 1.0.0).

    Releasing version 1.2 is also the blocker for merging a large feature
    branch on resource management that many contributors have worked on
    over the last months.

  - The community is also in the progress of releasing the bug fix release
    1.1.4.  This release is the result of a major effort with large
    production users to address robustness issues found in such
    installations.

  - There will be 'Flink Forward' conferences in San Francisco (April 2017)
    and Berlin (September 2017). The PMC will be in touch with the organizers
    to make sure the event is endorsed by the Flink community and respects
    the Apache trademark guidelines.

  - 'data Artisans' conducted a survey among Flink users and contributors,
    with the goal to share the results with the community.  The survey is
    currently being evaluated, results to be shared soon.

  - In the previous quarter, the Flink community started working together
    with the Apache Bahir community. The first connectors are now exclusively
    managed by the Bahir project, and not part of the Flink project any more.

  - Several community members have started to voice concerns that the Flink
    code base is getting very intensive to build and test. Individual
    discussions in these threads suggested to kick of a proper discussion
    after the 1.2 release is out.

COMMUNITY

There was no new PMC addition since the last report.
The latest PMC addition was July 2nd, 2015 (Maximilian Michels)

No new committers have been added since the last report.
The last committer addition was Tzu-Li (Gordon) Tai on August, 26th, 2016.

Flink currently has 24 committers and 16 PMC members.

RELEASES

The following releases were made since the last board report:

 → 1.1.3 was released on October, 12th, 2016

ACTIVITY ON JIRA / MAILING LISTS

Both user@f.a.o and dev@f.a.o lists are quite active, with mails getting
answered by a mix of committers and non-committers.

JIRA is even more active than in the three month period before
(730 JIRA tickets created, 545 JIRA tickets resolved)

21 Sep 2016 [Stephan Ewen / Mark]

DESCRIPTION

Apache Flink is a distributed data streaming system for batch and streaming
data analysis on top of a streaming dataflow engine.  Flink's stack contains
functional batch and streaming analysis APIs in Java and Scala and libraries
for various use cases.

Flink interacts and integrates with several Apache projects in the broader
ecosystem of data storage and computing, such as Apache Hadoop, Kafka, HBase,
Cassandra.

ISSUES

  - There are no issues that require board attention.

STATUS AND ACTIVITY

  - The Flink community started working together with the Apache Bahir
    community.  Some of the streaming connectors from Flink are now maintained
    under the Bahir umbrella. This allows for faster independent release
    cycles and more lightweight processes to add new connectors.

    We discussed as an alternative to create an independent subproject of
    Flink, but the community favored to work with Bahir. The Bahir community
    was very open to the idea and it looks like both committers and
    contributors are quite happy with the result.

  - The second edition of the "Flink Forward" conference took place in Berlin
    from September 12th to 14th. It featured a series of talks about use
    cases, applications, and in-depth technical discussions, as well as a
    training session. About 350 people attended.

    A good amount of Flink committers were at the conference as well and
    participated with talks and discussions about various aspects of the
    system.

  - The community adopted the "Improvement Proposal" concept from the Kafka
    community, under the name "FLIP" = "Flink Improvement Proposal", and uses
    the Wiki for that.

    Improvement proposals are proposed major changes with thorough design
    docs. Once discussed and approved (via mailing list), they are usually
    broken down into JIRA issues which are linked.

    We found that this complements JIRA and the mailing lists, because (1)
    there is one place where interested users and contributors can find an
    overview of all major additions/changes that have been proposed and are
    pending, and (2) the Wiki document sums up the final conceptual designs
    nicely while JIRA tracks the implementation aspects.

  - In the previous quarter, the Flink community introduces the concept of
    "shepherds".  Shepherds are experts/trackers of components, but have no
    authoritative power.

    The concept seems to not have discouraged anyone to look across component
    boundaries, and occasionally, it has helped for threads to be picked up.
    Overall, however, it has not been picked up as much as hoped for. Maybe it
    will be picked up better over time (it might simply take a while until
    everybody intuitively adopts such able concept). For now, it does not seem
    to have any downsides.

COMMUNITY

There was no new PMC addition since the last report.
The latest PMC addition was July 2nd, 2015 (Maximilian Michels)

Tzu-Li (Gordon) Tai was added as a committer on August, 26th, 2016.

Flink currently has 24 committers and 16 PMC members.

RELEASES

The following releases were made since the last board report:

 - 1.1.0 was released on August, 4th, 2016
 - 1.1.1 was released on August, 10th, 2016
 - 1.1.2 was released on September, 5th, 2016

ACTIVITY ON JIRA / MAILING LISTS

Development speed is fast, we are seeing more thorough use of
JIRA, more fine grained issues, more discussion.

15 Jun 2016 [Stephan Ewen / Bertrand]

DESCRIPTION

Apache Flink is a distributed data streaming system for batch and streaming
data analysis on top of a streaming dataflow engine.  Flink's stack contains
functional batch and streaming analysis APIs in Java and Scala and libraries
for various use cases.  Flink interacts and integrates with several Apache
projects in the broader ecosystem of technology around Apache Hadoop and
Apache Kafka.

ISSUES

 - There are no issues that require board attention.

STATUS AND ACTIVITY

 - Shane informed us that the "Apache Flink" trademark has been approved,
   but still needs to be officially issued.

 - Since the 1.0 release, the community growth seems to have accelerated.

 - As a response, the community decided to promote a "shepherd" concept
   within the community.  Every component (like API, distributed runtime,
   library, ...) has one or more shepherds, which are experts of that
   component, and act as people to contact for issues and questions that
   cannot be answered by the broad community.

   We decided to go with the name "shepherd", rather than "maintainer",
   because it is a non-authoritative role, but a guiding role.

   Shepherds are mostly committers, but also some non-committers volunteered.
   This looks like a nice way to identify prospects for future committers.

   The shepherds should also feel responsible for driving/moderating (not
   deciding) the resolution of community contributions.

   We are still in the phase of trying this out, hope to be able to report
   results in the next board report.

 - There will be a conference on Apache Flink and Stream Processing in
   September called "Flink Forward" (http://flink-forward.org/). The
   organizers are in touch with the PMC. The program committee will be five
   people from the Apache stream processing ecosystem.

 - Flink has been integrated with more Apache projects, like BigTop (done by
   the Flink and BigTop community) and Mahout (mostly done by the Mahout
   community).

 - Since the 1.0 release, the community brought measures in place to check
   binary backwards compatibility.

 - So far, the Flink community has accepted the majority of contributions,
   such as for connectors for integration with external systems.  We have,
   however, seen quite a few times the contributors were not interested in
   following up and maintaining the contribution, and a tough maintenance job
   was imposed on the community for barely used and unfamiliar code.

   As a results, there are discussions about encouraging contributors to
   create individual independent projects for such connectors, if there are
   no additional people in the community that would back the contribution.

COMMUNITY

There was no new PMC addition since the last report.

The latest PMC addition was July 2nd, 2015 (Maximilian Michels)
The latest committer addition was on March, 8th, 2016 (Greg Hogan)

Flink currently has 23 committers and 16 PMC members.

RELEASES

The following releases were made since the last board report:

 - 1.0.1 was released on Tue Apr 05 2016
 - 1.0.2 was released on Fri Apr 22 2016
 - 1.0.3 was released on Wed May 11 2016

ACTIVITY ON JIRA / MAILING LISTS

Development speed continues to be healthy.

 - 439 JIRA tickets created in the last 3 months
 - 318 JIRA tickets closed/resolved in the last 3 months

Compared to previous quarters, dev mailing list activity is increasing
a bit, but user mailing list activity is increasing a lot.

 - dev@flink.apache.org:
    - 381 subscribers (up 58 in the last 3 months):
    - 1396 emails sent to list (1176 in previous quarter)

 - user@flink.apache.org:
    - 533 subscribers (up 119 in the last 3 months):
    - 2211 emails sent to list (1368 in previous quarter)

16 Mar 2016 [Stephan Ewen / Greg]

DESCRIPTION

Flink is a distributed data streaming system for batch and streaming data
analysis on top of a streaming dataflow engine.
Flink's stack contains functional batch and streaming analysis APIs in
Java and Scala and libraries for various use cases.
Flink interacts and integrates with several Apache projects in the broader
ecosystem of technology around Apache Hadoop and Apache Kafka.


ISSUES

 - There are no issues that require board attention.


STATUS AND ACTIVITY

 - The recent months where dominated by the efforts to release Flink 1.0, which
   was officially announced by the ASF on March, 8th.
 - The 1.0 release was themed around backwards compatibility, and operational
   features.  A complex event processing library was added as well.
 - A roadmap for the first half of 2016 has been published by the community on
   the mailing lists end of December.
 - Flink seems to get a decent amount of interest, several upcoming conference
   talks (e.g., Strata, Hadoop Summit, Kafka Summit, OSCon, ApacheCon, and
   several Meetups)
 - The latest release (1.0.0) saw slightly fewer contributors (64 people) than
   the previous one (0.10.0 had 79 contributors), because some features and
   contributions were held off / delayed in order to establish the 1.0
   relevant tools (like compatibility checks).


COMMUNITY

 - There was no new PMC addition since the last report.
   The latest PMC addition was July 2nd, 2015 (Maximilian Michels)

 - There were two committers added since the last report
    - ChengXiang Li was added on January, 20th, 2016
    - Greg Hogan was added on March, 8th, 2016

Flink currently has 23 committers and 16 PMC members


RELEASES

The following releases were made since the last board report:

 - 0.10.2 was released on February, 10th, 2016
 - 1.0.0 was released on March, 8th, 2016


ACTIVITY ON JIRA / MAILING LISTS

Development speed continues to be healthy.

 - 440 JIRA tickets created in the last 3 months
 - 374 JIRA tickets closed/resolved in the last 3 months


Compared to previous quarters, there is a bit of a shift in mailing list
activity from the dev list to the users list.

 - dev@flink.apache.org:
    - 323 subscribers (up 68 in the last 3 months):
    - 1266 emails sent to list (1617 in previous quarter)

 - user@flink.apache.org:
    - 414 subscribers (up 97 in the last 3 months):
    - 1528 emails sent to list (1225 in previous quarter)

 - issues@flink.apache.org:
    - 54 subscribers (up 9 in the last 3 months):
    - 8794 emails sent to list (11290 in previous quarter)

16 Dec 2015 [Stephan Ewen / Chris]

DESCRIPTION

Flink is a distributed data streaming system for batch and streaming data
analysis on top of a streaming dataflow engine.
Flink's stack contains functional batch and streaming analysis APIs in
Java and Scala and libraries for various use cases.

Flink interacts and integrates with several Apache projects in the broader
ecosystem of technology around Hadoop.


ISSUES

 - There are no issues that require board attention.


STATUS AND ACTIVITY

 - The community has released version 0.10, which had a strong focus on
   hardening the data streaming capabilities and adding major features like
   master high availability.

 - The community is actively working towards a version 1.0 and discussing which
   APIs and components will be maintained backwards compatible.

 - There was a 2 day conference on Flink on Berlin in October, with
   community-selected talks, which had attracted around 250 participants.

 - The Flink community is in a good growing phase. The latest release had
   contributions from 79 people across many organizations.


COMMUNITY

 - The newest PMC member was added July 2nd, 2015 (Maximilian Michels)
 - The newest committer was added August 31st 2015 (Matthias J. Sax)

Flink currently has 21 committers and 16 PMC members


RELEASES

The following releases were made since the last board report:

 - 0.10.1 was released on Thu Nov 26 2015
 - 0.10.0 was released on Fri Nov 13 2015


ACTIVITY ON JIRA / MAILING LISTS

 - 506 JIRA tickets created in the last 3 months
 - 434 JIRA tickets closed/resolved in the last 3 months

 - dev@flink.apache.org:
    - 249 subscribers (up 35 in the last 3 months):
    - 1708 emails sent to list (1729 in previous quarter)

 - news@flink.apache.org:
    - 105 subscribers (up 18 in the last 3 months):
    - 1 emails sent to list (2 in previous quarter)

 - user@flink.apache.org:
    - 308 subscribers (up 73 in the last 3 months):
    - 1340 emails sent to list (1176 in previous quarter)

 - issues@flink.apache.org:
    - 42 subscribers (up 4 in the last 3 months):
    - 12105 emails sent to list (10466 in previous quarter)

16 Sep 2015 [Stephan Ewen / Chris]

DESCRIPTION

Flink is a distributed data streaming system for batch and streaming data
analysis on top of a streaming dataflow engine.
Flink's stack contains functional batch and streaming analysis APIs in
Java and Scala and libraries for various use cases.

Flink interacts and integrates with several Apache projects, including
Hadoop (YARN, HDFS, MapReduce), Kafka, HBase, Zeppelin, MRQL, Tez, Orc,
HCatalog.


ISSUES

 - There are no issues that require board attention.


STATUS AND ACTIVITY

 - The community is working towards version 0.10, which contains several new
   major features, like an overhaul of the streaming API, high availability.
   The community aims to graduate the streaming API out of beta status and make
   it stable.

 - The question about a releasing version 1.0 has been brought up in the
   community, as APIs and operational features mature.

 - Flink is seeing an increasing number of users and contributors, and the
   community is discussing ways to handle the load, make pull requests and
   patches more efficient to review, and to guide contributors towards
   making productive contributions. Current efforts include more precise
   guidelines and automated QA technology.

 - Interest in Flink is growing, we are seeing more and more meetup groups
   springing up, as well as a 2-day conference on Flink in October in Berlin.

 - The community is looking to diversify the technical expertise.
   Right now, most people contribute to libraries, APIs, tools, integration.
   We would like to get more people knowledgeable about the core. Efforts in
   talks and documentation are currently happening.


COMMUNITY

 - The project has one new PMC member since the last report
   (Maximilian Michels was added July 2nd, 2015)

 - The project has 4 new committers since the last board report

   - Andra Lungu was added as a committer on June 08th, 2015
    (voted in before last report, accepted and added since)
   - Chiwan Park was added as a committer on June 22 2015
   - Chesnay Schepler was added as a committer on Tue, August 11th, 2015
   - Matthias J. Sax was added as a committer on Mon, August 31st 2015

 - Flink currently has 21 committers and 16 PMC members


RELEASES

The following releases were made since the last board report

 - 0.9.0 was released on June 24th, 2015
 - 0.9.1 was released on August 31st, 2015


MAILING LISTS

 - dev@flink.apache.org:
    - 214 subscribers (up 40 in the last 3 months):
    - 1661 emails sent to list (1956 in previous quarter)

 - user@flink.apache.org:
    - 233 subscribers (up 74 in the last 3 months):
    - 1183 emails sent to list (756 in previous quarter)

 - issues@flink.apache.org:
    - 40 subscribers (up 2 in the last 3 months):
    - 10271 emails sent to list (8179 in previous quarter)


JIRA ACTIVITY

 - 446 JIRA tickets created in the last 3 months
 - 350 JIRA tickets closed/resolved in the last 3 months

17 Jun 2015 [Stephan Ewen / Brett]

DESCRIPTION

Flink is a distributed data analytics system for batch and streaming data
analysis on top of a streaming dataflow engine.
Flink's stack contains functional batch and streaming analysis APIs in
Java and Scala and libraries for various use cases.
Flink is a citizen of the Hadoop ecosystem and interacts
with HDFS, YARN, MapReduce, HBase, and Tez.


ISSUES

There are no issues that require board attention.


STATUS

The Flink community is preparing the release of version 0.9, which marks
the biggest additions to Flink since the project creation. The system
is adding strong functionality for stream processing and libraries
for various use cases.

Flink has integrated with various other Apache projects (Tez, Zeppelin, Samoa).

The community is growing, we see increasingly articles and meetups on Flink
appearing.


COMMUNITY

* Last committer was voted in May 30th, 2015 (Andra Lungu)

* Vote to add Chiwan Park as committer is happening as of writing this report.

* Last PMC member was added December 17, 2014, when all committers and a subset
  of the mentors formed the project Flink PMC during graduation.

* Project has currently 18 committers and 15 PMC members.

* The project community is active (May: 274 messages in user mailing list,
  491 messages in dev mailing list, 2758 messages in issues mailing list)

* JIRA reports 194 new issues and 152 solved issues in the last 30 days


COMMUNITY OBJECTIVES

* Growing the user community further via talks and meetups and integration with
  other projects. The Flink PMC plans to organize a track at the Apache Big
  Data conference.

* Integration with other Apache projects is continuously happening:
  - Apache Zeppelin (incubating) has added support for Flink
  - Apache Samoa (incubating) integration is a pending patch
  - Apache Tez integration is part of the next release.
  - Integration with ORC and HCatalogue are work in progress.

RELEASES

 * Date of last release: April 13th, 2015 (version 0.9.0-milestone-1)
 * Release candidate creation for 0.9.0 is happening as of writing this report.

18 Mar 2015 [Stephan Ewen / Doug]

DESCRIPTION

Flink is a distributed data analytics system for expressive, declarative,
and efficient batch and streaming data processing and analysis.
Flink's stack contains functional batch and streaming analysis APIs in
Java and Scala, a type analysis and serialization framework, a library of
out-of-core enabled operators, distributed task scheduling, and a data
streaming network subsystem.
Flink is a citizen of the Hadoop ecosystem and interacts
with HDFS, YARN, MapReduce, HBase, and Tez.


ISSUES

There are no issues that require board attention.


STATUS

The Flink community is working towards to next version (0.9) which
builds Flink into a proper hybrid batch/streaming system with hardened
guarantees.

The community is growing, we see increased interest in contributions
from companies that were not previously involved with Flink.


COMMUNITY

* Last committer was voted in March 5th, 2015 (Paris Carbone)

* Last PMC member was added December 17, 2014, when all committers and a subset
  of the mentors formed the project Flink PMC during graduation.

* Project has currently 17 committers and 15 PMC members.

* The project community is active (February: 123 messages in user mailing list,
  545 messages in dev mailing list, 1980 messages in issues mailing list)

* JIRA reports 185 new issues and 120 solved issues in the last 30 days


COMMUNITY OBJECTIVES

* Growing the user community further. The community will give talks at
  various meetups and conferences, including Hadoop Summit, Strata, ApacheCon.

* Integration with other Apache projects is work in progress.
  Patches that integrate Flink with Tez and SAMOA are available.


RELEASES

 * Date of last release: February 19th, 2015 (version 0.8.1)

18 Feb 2015 [Stephan Ewen / Rich]

DESCRIPTION

Flink is a distributed Big Data system for expressive, declarative,
and efficient batch and streaming data processing and analysis.
Flink’s stack contains functional batch and streaming analysis APIs in
Java and Scala, a type analysis and serialization framework, a library of
out-of-core enabled operators, distributed task scheduling, and a data
streaming network subsystem.
Flink is a citizen of the Hadoop ecosystem and interacts
with HDFS, YARN, MapReduce, HBase, and Tez.


ISSUES

There are no issues that require board attention.


STATUS

Flink graduated in December from the Incubator. The transition is complete
by now, the first release and committer addition as a TLP have happened.

The community is growing, judging from mailing list activity, number of
meetup group talks, bug reports, etc


COMMUNITY

* Last committer was voted in February 6th, 2015 (Maximilian Michels)

* Last PMC member was added December 17, 2014, when all committers and a subset
  of the mentors formed the project Flink PMC during graduation.

* The project community is active (February 1st to 11th: 57 messages in user
  mailing list, 191 messages in dev mailing list, 244 messages in commits
  list, 887 messages in issues mailing list)

* Community is actively developing the software. JIRA reports 139 new
  issues and 121 solved issues in the last 30 days

* The issue of trademarking Flink is still ongoing. The process has started
  (same as last month)


COMMUNITY OBJECTIVES

* Growing the user community further

* Integration with other Apache projects is work in progress.
  As of the last 30 days specifically SAMOA (incubating) and Tez (per mailing
  lists).


RELEASES

 * Date of last release: January 19th, 2015 (version 0.8.0)
 * Candidate for release version 0.8.1 has been created (February 10th).

21 Jan 2015 [Stephan Ewen / Sam]

DESCRIPTION
-----------
Flink is a distributed Big Data system for expressive, declarative,
and efficient batch and streaming data processing and analysis.

Flink’s stack contains functional batch and streaming analysis APIs in
Java and Scala, a type analysis and serialization framework, a library of
out-of-core enabled operators, distributed task scheduling, and a data
streaming network subsystem.

Flink is a citizen of the Hadoop ecosystem and interacts with HDFS, YARN,
MapReduce, HBase, and Tez.


ISSUES
------
There are no issues that require board attention.


STATUS
------
Flink recently graduated in December from the Incubator. The infrastructure
for becoming a top-level project has been set up, the renaming of the mailing
lists is partially done (as of the time of writing this report).

The Apache press team announced the graduation on January 12th. There is a
decent media echo.


COMMUNITY
---------
* Last committer was voted in November 20th, 2014 (Timo Walther)

* When graduating to a TLP (December 17, 2014), all committers and
  a subset of the mentors formed the project Flink PMC

* The project community is active and growing
 (January 1st to 11th: 25 messages in user mailing list,
  180 messages in dev mailing list, 126 messages in commits list,
  314 messages in issues mailing list)

* Community is actively developing the software.
  JIRA reports 72 new issues and 54 solved issues in the last 30 days

* The issue of trademarking Flink is still ongoing. The process has started.

* December and January have various Flink-related meetups (sample below):
 - December 11th – Netherlands Hadoop User Group
 - January 21st – Bay Area Hadoop User Group
 - January 28 – Bay Area Tez User Group
 - January 28th – Berlin Flink User Group


COMMUNITY OBJECTIVES
--------------------
* Defining and agreeing upon a new 6-12 months roadmap (currently discussed
  on the mailing lists)

* Growing the user community further

* As a special case of the community growth, Flink would like to interact
  stronger with other Apache projects (TLP and Incubator). Roadmap draft
  includes integrating for example with Zeppelin, SAMOA, Tez, and Mahout.
  Integration with Hive is under discussion.


RELEASES
--------
* Date of last release: Oct 26th (0.7-incubating),
  while Flink was still in the incubator.

* Release candidates for version 0.8 have been created
  (RC1 rejected by PMC vote, RC2 validation happening as of the
   time of writing this report)

17 Dec 2014

Establish the Apache Flink Project

 WHEREAS, the Board of Directors deems it to be in the best
 interests of the Foundation and consistent with the
 Foundation's purpose to establish a Project Management
 Committee charged with the creation and maintenance of
 open-source software, for distribution at no charge to
 the public, related to fast and reliable large-scale data
 analysis with focus on programmability, optimizability,
 efficiency, and combination of batch and streaming data
 processing.

 NOW, THEREFORE, BE IT RESOLVED, that a Project Management
 Committee (PMC), to be known as the "Apache Flink Project",
 be and hereby is established pursuant to Bylaws of the
 Foundation; and be it further

 RESOLVED, that the Apache Flink Project be and hereby is
 responsible for the creation and maintenance of software
 related to fast and reliable large-scale data analysis
 with focus on programmability, optimizability, efficiency,
 and combination of batch and streaming data processing;
 and be it further

 RESOLVED, that the office of "Vice President, Apache Flink" be
 and hereby is created, the person holding such office to
 serve at the direction of the Board of Directors as the chair
 of the Apache Flink Project, and to have primary responsibility
 for management of the projects within the scope of
 responsibility of the Apache Flink Project; and be it further

 RESOLVED, that the persons listed immediately below be and
 hereby are appointed to serve as the initial members of the
 Apache Flink Project:

    * Márton Balassi            <mbalassi@apache.org>
    * Ufuk Celebi               <uce@apache.org>
    * Stephan Ewen              <sewen@apache.org>
    * Gyula Fora                <gyfora@apache.org>
    * Alan Gates                <gates@apache.org>
    * Fabian Hueske             <fhueske@apache.org>
    * Vasiliki Kalavri          <vasia@apache.org>
    * Aljoscha Krettek          <aljoscha@apache.org>
    * Robert Metzger            <rmetzger@apache.org>
    * Till Rohrmann             <trohrmann@apache.org>
    * Henry Saputra             <hsaputra@apache.org>
    * Sebastian Schelter        <ssc@apache.org>
    * Kostas Tzoumas            <ktzoumas@apache.org>
    * Timo Walther              <twalthr@apache.org>
    * Daniel Warneke            <warneke@apache.org>

 NOW, THEREFORE, BE IT FURTHER RESOLVED, that Stephan Ewen
 be appointed to the office of Vice President, Apache Flink, to
 serve in accordance with and subject to the direction of the
 Board of Directors and the Bylaws of the Foundation until
 death, resignation, retirement, removal or disqualification,
 or until a successor is appointed; and be it further

 RESOLVED, that the initial Apache Flink PMC be and hereby is
 tasked with the creation of a set of bylaws intended to
 encourage open development and increased participation in the
 Apache Flink Project; and be it further

 RESOLVED, that the Apache Flink Project be and hereby
 is tasked with the migration and rationalization of the Apache
 Incubator Flink podling; and be it further

 RESOLVED, that all responsibilities pertaining to the Apache
 Incubator Flink podling encumbered upon the Apache Incubator
 Project are hereafter discharged.

 Special Order 7B, Establish the Apache Flink Project , was
 approved by Unanimous Vote of the directors present.

19 Nov 2014

Flink is an open source system for expressive, declarative, fast, and
efficient data analysis. Flink combines the scalability and programming
flexibility of distributed MapReduce-like platforms with the efficiency, out-
of-core execution, and query optimization capabilities found in parallel
databases. Flink was originally known as Stratosphere when it entered the
Incubator.

Flink has been incubating since 2014-04-14.

Three most important issues to address in the move towards graduation:

 1. Community growth


Any issues that the Incubator PMC (IPMC) or ASF Board wish/need to be
aware of?

  None

How has the community developed since the last report?

 Two new committers were voted into the project: Márton Balassi and Gyula
 Fóra. There were presentations on Flink at the VLDB keynote, a meetup in
 Stockholm (more than 150 participants), one in Pasadena, CA and two
 meetups in Berlin.  There are upcoming talks in Palo Alto, CA and at the
 Apache Con in Budapest. The dev@flink mailing list had more than 300
 messages on average in the last months. The user@ list 124 in October.

How has the project developed since the last report?

 Flink has released two major versions (0.6-incubating and 0.7-incubating)
 in the last three months. (There was also a minor bugfix release for the
 0.6-incubating release). The Apache MRQL (incubating) project has added
 support for running programs with Flink. There is an effort within the
Flink community to integrate with Apache Tez.

Date of last release:

 2014-10-26

When were the last committers or PMC members elected?

  2014-09-06


Signed-off-by:

 [ ](flink) Sean Owen
 [x](flink) Ted Dunning
 [ ](flink) Owen O'Malley
 [X](flink) Henry Saputra
 [ ](flink) Ashutosh Chauhan
 [X](flink) Alan Gates

20 Aug 2014

Flink is an open source system for expressive, declarative, fast, and
efficient data analysis. Flink combines the scalability and programming
flexibility of distributed MapReduce-like platforms with the efficiency,
out-of-core execution, and query optimization capabilities found in parallel
databases. Flink was originally known as Stratosphere when it entered the
Incubator.

Flink has been incubating since 2014-04-14.

Three most important issues to address in the move towards graduation:

 1. First incubator release
 2. Continue with community growth
 3. More comfortable with ASF policies and processes

Any issues that the Incubator PMC (IPMC) or ASF Board wish/need to be
aware of?

 None

How has the community developed since the last report?

 One new committer has been accepted into the project.
 We received a large external code contribution and pending reviews.
 The project started a bi-weekly Hangout.

How has the project developed since the last report?

 The project has moved all infrastructure to Apache,
 the source code has been renamed to the "org.apache.flink" namespace.
 We are finalizing the preparations for the first Incubator release

Date of last release:

 2014-05-31 (no incubator release yet)

When were the last committers or PMC members elected?

 2014-07-10

Signed-off-by:

 [ ](flink) Sean Owen
 [ ](flink) Ted Dunning
 [ ](flink) Owen O'Malley
 [X](flink) Henry Saputra
 [ ](flink) Ashutosh Chauhan

16 Jul 2014

Flink is an open source system for expressive, declarative, fast, and
efficient data analysis. Flink combines the scalability and programming
flexibility of distributed MapReduce-like platforms with the efficiency,
out-of-core execution, and query optimization capabilities found in parallel
databases. Flink was originally known as Stratosphere when it entered the
Incubator.

Flink has been incubating since 2014-04-14.

Three most important issues to address in the move towards graduation:

 1. The new name ("Flink") has still not been confirmed by the trademark
    team.
 2. Setup of release infrastructure and first release
 3. Continue with community growth

Any issues that the Incubator PMC (IPMC) or ASF Board wish/need to be aware
of?

 None.

How has the community developed since the last report?

 Two presentations (Meetup in Nice by a committer, one in Budapest by a
 contributor), new users are showing up on the mailing list and on JIRA.

How has the project developed since the last report?

 Most of the infrastructure setup has been done. Users start using the new
 mailing lists.
 In the last month, 21 authors have pushed 120 commits to master. On
 master, 910 files have changed and there have been 74,313 additions and
 37,094 deletions.

Date of last release:

 2014-05-31 (no incubator release yet)

When were the last committers or PMC members elected?

 None.


Signed-off-by:

 [ ](flink) Sean Owen
 [ ](flink) Ted Dunning
 [ ](flink) Owen O'Malley
 [X](flink) Henry Saputra
 [ ](flink) Ashutosh Chauhan
 [X](flink) Alan Gates

18 Jun 2014

Flink is an open source system for expressive, declarative, fast, and
efficient data analysis. Flink combines the scalability and programming
flexibility of distributed MapReduce-like platforms with the efficiency,
out- of-core execution, and query optimization capabilities found in
parallel databases.

Flink has been incubating since 2014-04-14.

Three most important issues to address in the move towards graduation:

 1. Infrastructure setup (git, JIRA, mailing list renaming)
 2. Release
 3. Build an ASF community

Any issues that the Incubator PMC (IPMC) or ASF Board wish/need to be
aware of?

 None.

How has the community developed since the last report?

 The community still "lives" outside the ASF since the infrastructure is
 not ready.

 Two new contributors opened pull requests since the last report.

How has the project developed since the last report?

 The community has voted for a new name, so the project is moving away from
 its old name "Stratosphere" to the new one: "Flink".

 We are still waiting for Infra to rename the mailing list, to create a
 JIRA (blocked by the mailing list) and create a Git repository (blocked by
 the SGA (which we already provided)).

 We released a new major release (0.5) outside the ASF.

Date of last release:

 2014-05-31 (no incubator release yet)

When were the last committers or PMC members elected?

 Only initial committers.

Signed-off-by:

 [ ](flink) Sean Owen
 [X](flink) Ted Dunning
 [ ](flink) Owen O'Malley
 [X](flink) Henry Saputra
 [ ](flink) Ashutosh Chauhan
 [X](flink) Alan Gates

Shepherd/Mentor notes:

 Konstantin Boudnik (cos):

   Project dev@ list has healthy traffic.  Some of the mentors are active
   on the list and helping community. No visible issues.

21 May 2014

Stratosphere is an open source system for expressive, declarative, fast, and
efficient data analysis. Stratosphere combines the scalability and programming
flexibility of distributed MapReduce-like platforms with the efficiency, out-
of-core execution, and query optimization capabilities found in parallel
databases.

Stratosphere has been incubating since 2014-04-14.

Three most important issues to address in the move towards graduation:

 1. Find a new project name
 2. Release
 3. Build an ASF community

Any issues that the Incubator PMC (IPMC) or ASF Board wish/need to be
aware of?

 None.

How has the community developed since the last report?

 This is the first report. We got 2 Google Summer of Code students
 and one new external contributor opened a pull request.

How has the project developed since the last report?

 This is the first report. We are stabilizing the code base
 for a new major release (0.5).
 We are currently setting up the initial infrastructure at
 the ASF incubator.

Date of last release:

 2014-01-13 (no incubator release yet)

When were the last committers or PMC members elected?

 Only initial committers.

Signed-off-by:

 [ ](stratosphere) Sean Owen
 [x](stratosphere) Ted Dunning
 [ ](stratosphere) Owen O'Malley
 [ ](stratosphere) Henry Saputra
 [ ](stratosphere) Ashutosh Chauhan
 [X](stratosphere) Alan Gates

Shepherd/Mentor notes:

  Dev. list look pretty good, traffic is high and mentors participation is
  certainly there.