You are viewing a plain text version of this content. The canonical link for it is here.
Posted to general@incubator.apache.org by Emmanuel Lecharny <el...@gmail.com> on 2010/01/12 13:24:49 UTC
[Proposal] JPPF : a parallel processing framework for Java
Hi guys,
On behalf of the JPPF community, I'd like to bring the following
proposal for discussion within the Incubator. I had a meeting
with Laurent last week when he investigated about using MINA into
JPPF, and I suggested that this project could be a good fit for
the Incubator.
We will need mentors for this proposal
WDYT ?
Thanks!
http://wiki.apache.org/incubator/JppfProposal
--- Current Wiki Text below ---
= JPPF : a parallel processing framework for Java =
=== Abstract ===
JPPF enables applications with large processing power requirements to be
run on any number of computers, in order to dramatically reduce their
processing time. This is done by splitting an application into smaller
parts that can be executed simultaneously on different machines.
=== Proposal ===
JPPF aims at facilitating the parallelization of computation-intensive
applications, with a focus on ease of use, performance and reliability.
To achieve these goals, the framework comes with a number of outstanding
features:
- Comprehensive, easy to use APIs: passing from a single-threaded
application model to a grid-based parallel model can be a daunting task.
JPPF facilitates this work by providing developers with a set of APIs
that are simple, can be learned quickly and require a minimal or no
modification to the existing code.
- No configuration usage: in most environments, JPPF can be deployed
without any additional configuration burden. Nodes and application
clients will automatically dicover the servers on the network. The
server will automatically adapt to workload changes and optimize the
throughput. Required code and libraries will be automatically deployed
where they are needed.
- Dynamic grid scaling and self-repair: the JPPF grid is fault-tolerant,
meaning that the failure of a node, or even a server, does not
compromise the jobs currently executing or scheduled. In most cases, the
performance degradation will be barely noticeable, as JPPF automatically
adapts to topology and workload changes. Furthermore, nodes and servers
can be dynamically started and will be automatically recognized,
allowing JPPF to function in "crunch mode". In addition to this, JPPF
components benefit from automatic recovery functionalities.
- Job-level SLA: each job submitted to the JPPF grid runs within limits
defined by its own SLA (service level agreement). This allows to specify
the characteristics (i.e. available memory, processors, disk space,
operating systems, etc.) of the nodes a job can run on, as well as how
many nodes it can run on. As many functionalities in JPPF, this one can
be dynamically adjusted, manually or automatically.
- Advanced Management and monitoring: full-fledged management and
monitoring features are provided out of the box: server and nodes status
monitoring, detailed statistics and events, remote administration,
job-level real-time monitoring and management, charts, cpu utilization
(for billing). These functionalities are available via a graphical user
interface as well as from the JPPF APIs.
- Integration with leading application and web servers: by complying
with the Java Connector Architecture 1.5 specification, JPPF integrates
seamlessly with and completes the offering of leading J2EE application
servers: Apache Geronimo, JBoss, Glassfish, IBM Websphere, Oracle
Weblogic, Oracle OC4J. JPPF also integrates with GigaSpaces eXtreme
Application Platform and Apache Tomcat web server
- Extensibility: JPPF offers a number of hooks and extension points that
allow users to extend the framework and adapt it to their own
requirements and needs. Such extension mechanisms are available for
custom management and monitoring MBeans, startup classes for grod
components, network data transformation/encryption, additional
load-balancing algorithms, alternate object serialization mechanisms.
=== Background ===
JPPF is a project being actively developed at SourceForge. It was
created to address a class of problems called "embarassingly parallel",
which groups computational problems that can be decomposed into many
smaller sub-problems, that are independant from each other and that can
thus be executed in parallel.
=== Rationale ===
Given these last years' emergence of technologies that make commodity
hardware, virtualization and cloud computing available to a fast-growing
computing ecosystem, the project answers the need to execute
applications ever faster, with a low entry cost, while at the same time
preserving historical technological investments.
== Current Status ==
=== Meritocracy ===
MWe acknowledge that a meritocratic governance is the only way for the
project to grow and expand, in the spirit of open source and the ASF. It
will benefit the project, its communities, the ASF, as well as the outer
ecosystems.
=== Community ===
JPPF already has a relatively small, but steadily growing community of
users. Given the applicability of the project to numerous industries and
technological areas such as scientific research, finance,
graphical/video rendering, telecoms, data mining, etc.., we are
confident that there is a very large growth potential for developers and
users communities around JPPF.
=== Core Developers ===
JPPF was founded in April 2005 by Laurent Cohen (laurent.cohen at
jppf.org), who is currently the only active developer and code committer.
The other active contributor is John Channing (john.channing at
gmail.com) who provides advice and peer review on the project's
architecture, design, requirements and promotion.
Other project members and former project contributors include:
- Domingos Creado (dcreado at users.sourceforge.net)
- Mahendra Kutare (mahendra.kutare at gmail.com)
- Guy Korland (Guy at gigaspaces.com)
- Fredreric Barachant (pepe-barachant at users.sourceforge.net)
- Peter Becker (nc-heuelhe at netcologne.de)
- Wolfgang Wagner (wolfgang.wagner at iname.com)
- Jay Yusko (jay.yusko at gensolco.com)
=== Alignment ===
The project is currently undergoing a refactoring of its network
communication infrastructure, relying essentially on Mina, in order to
provide a greater maintainability of the code, as well as extend its
scalability and provide new functionalities such as truly secure
communications and integration of other distributed computing models
(e.g. P2P, map/reduce).
JPPF also offers connectors for Geronimo and Tomcat.
== Known Risks ==
=== Orphaned products ===
This is the main and most obvious risk of this project. The knowledge of
the code is owned by a single developer, Laurent Cohen, who has
committed about 98% of the existing code in the current repository. We
are well aware of this problem, and it is our hope and challenge that
integrating the Apache community will inspire the growth of a strong
developers community around JPPF.
=== Inexperience with Open Source ===
JPPF was created as an Open Source project in April 2005, and has
remained so since.
=== Homogenous Developers ===
A single active developer implies homogeneity.
=== Reliance on Salaried Developers ===
ANo one is currently paid to work on JPPF.
=== Relationships with Other Apache Products ===
JPPF is currently refactoring its network communication infrastructure
with the help of Mina (which is how we came to find a sponsor).
The project's logging relies entirely on commons-logging and Log4j.
The project also provides connectors for Geronimo and Tomcat.
Builds are done with Ant, and we are currently looking at switching to
Maven-based builds.
We foresee that JPPF could be used by many other Apache projects to
speed-up the execution of unit tests by running them in parallel, and to
serve as a basis for running distributed charge and load tests.
One of our samples uses Lucene to demonstrate a simple use of a
distributed framework for web search, crawling and indexing
=== A Excessive Fascination with the Apache Brand ===
Our main hope in joining the Apache community is that it will help build
a strong developers and committers community around the project, and
remove its reliance on a single developer. We aim at achieving this
while attracting new users, in conformance with the Apache spirit and
policies.
== Documentation ==
Information on JPPF can be found at:
http://www.jppf.org (main site, documentation and user forums)
http://sourceforge.net/projects/jppf-project (code repository, bug
tracker, features tracker)
== Initial Source ==
OThe entire code for JPPF is held in a SourceForge CVS repository, and
has been so since the project's inception in April 2005. This entire
code base will be donated to Apache.
The source code has been licensed under the ASL 2.0 since August 2007.
Before that it was LGPL.
All artifacts in the trunk and existing branches (1 branch) are now
licensed under the ASL 2.0, with corresponding header when applicable.
== External Dependencies ==
Other than the Apache products used in JPPF, the project currently
depends on the following ASL 2.0 compatible products/licenses:
- Rhino (MPL)
- JGoodies Looks (BSD)
- Groovy (ASL 2.0)
- Hazelcast (ASL 2.0)
- JUnit (CPL)
- MigLayout (BSD)
- Smart and Simple Web Crawler (ASL 2.0)
Dependencies on libraries with non AL-compatible licenses:
- JFreeChart (LGPL) - runtime dependency only
- SaverBeans SDK (LGPL) - in the process of being removed
- JAligner (GPL)
- Izpack (GPL)
- NSIS (zlib/libpng, bzip2, CPL) - in the process of being removed
== Cryptography ==
JPPF does not have any cryptographic component. However, the
distribution includes a sample that shows encryption/decryption of data
as a demonstration of one of its features. The sample is delivered with
full source code and can be found at:
http://www.jppf.org/wiki/index.php?title=Extending_and_Customizing_JPPF#Transforming_and_encrypting_networked_data
== Required Resources ==
Mailing lists
* jppf-private (with moderated subscriptions)
* jppf-dev
* jppf-commits
* jppf-user
Subversion Repository
* https://svn.apache.org/repos/asf/incubator/jppf
Issue Tracking
* JIRA (JPPF)
Others
* Web site: Confluence (JPPF)
== Initial Committers ==
* Laurent Cohen (laurent.cohen at jppf.org)
* John Channing (john.channing at gmail.com)
== Affiliations ==
None of the initial committers are paid by their employer, nor do they
represent their employer in any activity related to JPPF.
== Sponsors ==
Champion
* Emmanuel Lecharny (elecharny at apache dot org)
Nominated Mentors
We are currently looking for mentors within the community.
Sponsoring Entity
* Apache Incubator
--
Regards,
Cordialement,
Emmanuel Lécharny
www.nextury.com
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: [Proposal] JPPF : a parallel processing framework for Java
Posted by Jeremias Maerki <de...@jeremias-maerki.ch>.
A great idea! I like JPPF's approach whereas I've always had trouble
really grasping the map-reduce approach at grid computing (I'm not
coming from the search engine side).
I'm a bit reluctant to sign up as a mentor just yet because I'm still in
a very busy period and probably couldn't dedicate enough time right now.
Otherwise, I would have signed up immediately.
On 12.01.2010 13:24:49 Emmanuel Lecharny wrote:
> Hi guys,
>
> On behalf of the JPPF community, I'd like to bring the following
> proposal for discussion within the Incubator. I had a meeting
> with Laurent last week when he investigated about using MINA into
> JPPF, and I suggested that this project could be a good fit for
> the Incubator.
>
> We will need mentors for this proposal
>
> WDYT ?
>
> Thanks!
>
> http://wiki.apache.org/incubator/JppfProposal
>
> --- Current Wiki Text below ---
> = JPPF : a parallel processing framework for Java =
> === Abstract ===
> JPPF enables applications with large processing power requirements to be
> run on any number of computers, in order to dramatically reduce their
> processing time. This is done by splitting an application into smaller
> parts that can be executed simultaneously on different machines.
>
> === Proposal ===
> JPPF aims at facilitating the parallelization of computation-intensive
> applications, with a focus on ease of use, performance and reliability.
>
> To achieve these goals, the framework comes with a number of outstanding
> features:
>
> - Comprehensive, easy to use APIs: passing from a single-threaded
> application model to a grid-based parallel model can be a daunting task.
> JPPF facilitates this work by providing developers with a set of APIs
> that are simple, can be learned quickly and require a minimal or no
> modification to the existing code.
>
> - No configuration usage: in most environments, JPPF can be deployed
> without any additional configuration burden. Nodes and application
> clients will automatically dicover the servers on the network. The
> server will automatically adapt to workload changes and optimize the
> throughput. Required code and libraries will be automatically deployed
> where they are needed.
>
> - Dynamic grid scaling and self-repair: the JPPF grid is fault-tolerant,
> meaning that the failure of a node, or even a server, does not
> compromise the jobs currently executing or scheduled. In most cases, the
> performance degradation will be barely noticeable, as JPPF automatically
> adapts to topology and workload changes. Furthermore, nodes and servers
> can be dynamically started and will be automatically recognized,
> allowing JPPF to function in "crunch mode". In addition to this, JPPF
> components benefit from automatic recovery functionalities.
>
> - Job-level SLA: each job submitted to the JPPF grid runs within limits
> defined by its own SLA (service level agreement). This allows to specify
> the characteristics (i.e. available memory, processors, disk space,
> operating systems, etc.) of the nodes a job can run on, as well as how
> many nodes it can run on. As many functionalities in JPPF, this one can
> be dynamically adjusted, manually or automatically.
>
> - Advanced Management and monitoring: full-fledged management and
> monitoring features are provided out of the box: server and nodes status
> monitoring, detailed statistics and events, remote administration,
> job-level real-time monitoring and management, charts, cpu utilization
> (for billing). These functionalities are available via a graphical user
> interface as well as from the JPPF APIs.
>
> - Integration with leading application and web servers: by complying
> with the Java Connector Architecture 1.5 specification, JPPF integrates
> seamlessly with and completes the offering of leading J2EE application
> servers: Apache Geronimo, JBoss, Glassfish, IBM Websphere, Oracle
> Weblogic, Oracle OC4J. JPPF also integrates with GigaSpaces eXtreme
> Application Platform and Apache Tomcat web server
>
> - Extensibility: JPPF offers a number of hooks and extension points that
> allow users to extend the framework and adapt it to their own
> requirements and needs. Such extension mechanisms are available for
> custom management and monitoring MBeans, startup classes for grod
> components, network data transformation/encryption, additional
> load-balancing algorithms, alternate object serialization mechanisms.
>
> === Background ===
> JPPF is a project being actively developed at SourceForge. It was
> created to address a class of problems called "embarassingly parallel",
> which groups computational problems that can be decomposed into many
> smaller sub-problems, that are independant from each other and that can
> thus be executed in parallel.
>
> === Rationale ===
> Given these last years' emergence of technologies that make commodity
> hardware, virtualization and cloud computing available to a fast-growing
> computing ecosystem, the project answers the need to execute
> applications ever faster, with a low entry cost, while at the same time
> preserving historical technological investments.
>
> == Current Status ==
> === Meritocracy ===
> MWe acknowledge that a meritocratic governance is the only way for the
> project to grow and expand, in the spirit of open source and the ASF. It
> will benefit the project, its communities, the ASF, as well as the outer
> ecosystems.
>
> === Community ===
> JPPF already has a relatively small, but steadily growing community of
> users. Given the applicability of the project to numerous industries and
> technological areas such as scientific research, finance,
> graphical/video rendering, telecoms, data mining, etc.., we are
> confident that there is a very large growth potential for developers and
> users communities around JPPF.
>
> === Core Developers ===
> JPPF was founded in April 2005 by Laurent Cohen (laurent.cohen at
> jppf.org), who is currently the only active developer and code committer.
> The other active contributor is John Channing (john.channing at
> gmail.com) who provides advice and peer review on the project's
> architecture, design, requirements and promotion.
> Other project members and former project contributors include:
> - Domingos Creado (dcreado at users.sourceforge.net)
> - Mahendra Kutare (mahendra.kutare at gmail.com)
> - Guy Korland (Guy at gigaspaces.com)
> - Fredreric Barachant (pepe-barachant at users.sourceforge.net)
> - Peter Becker (nc-heuelhe at netcologne.de)
> - Wolfgang Wagner (wolfgang.wagner at iname.com)
> - Jay Yusko (jay.yusko at gensolco.com)
>
> === Alignment ===
> The project is currently undergoing a refactoring of its network
> communication infrastructure, relying essentially on Mina, in order to
> provide a greater maintainability of the code, as well as extend its
> scalability and provide new functionalities such as truly secure
> communications and integration of other distributed computing models
> (e.g. P2P, map/reduce).
> JPPF also offers connectors for Geronimo and Tomcat.
>
> == Known Risks ==
> === Orphaned products ===
> This is the main and most obvious risk of this project. The knowledge of
> the code is owned by a single developer, Laurent Cohen, who has
> committed about 98% of the existing code in the current repository. We
> are well aware of this problem, and it is our hope and challenge that
> integrating the Apache community will inspire the growth of a strong
> developers community around JPPF.
>
> === Inexperience with Open Source ===
> JPPF was created as an Open Source project in April 2005, and has
> remained so since.
>
> === Homogenous Developers ===
> A single active developer implies homogeneity.
>
> === Reliance on Salaried Developers ===
> ANo one is currently paid to work on JPPF.
>
> === Relationships with Other Apache Products ===
> JPPF is currently refactoring its network communication infrastructure
> with the help of Mina (which is how we came to find a sponsor).
> The project's logging relies entirely on commons-logging and Log4j.
> The project also provides connectors for Geronimo and Tomcat.
> Builds are done with Ant, and we are currently looking at switching to
> Maven-based builds.
> We foresee that JPPF could be used by many other Apache projects to
> speed-up the execution of unit tests by running them in parallel, and to
> serve as a basis for running distributed charge and load tests.
> One of our samples uses Lucene to demonstrate a simple use of a
> distributed framework for web search, crawling and indexing
>
> === A Excessive Fascination with the Apache Brand ===
> Our main hope in joining the Apache community is that it will help build
> a strong developers and committers community around the project, and
> remove its reliance on a single developer. We aim at achieving this
> while attracting new users, in conformance with the Apache spirit and
> policies.
>
> == Documentation ==
> Information on JPPF can be found at:
> http://www.jppf.org (main site, documentation and user forums)
> http://sourceforge.net/projects/jppf-project (code repository, bug
> tracker, features tracker)
>
> == Initial Source ==
> OThe entire code for JPPF is held in a SourceForge CVS repository, and
> has been so since the project's inception in April 2005. This entire
> code base will be donated to Apache.
> The source code has been licensed under the ASL 2.0 since August 2007.
> Before that it was LGPL.
> All artifacts in the trunk and existing branches (1 branch) are now
> licensed under the ASL 2.0, with corresponding header when applicable.
>
> == External Dependencies ==
> Other than the Apache products used in JPPF, the project currently
> depends on the following ASL 2.0 compatible products/licenses:
> - Rhino (MPL)
> - JGoodies Looks (BSD)
> - Groovy (ASL 2.0)
> - Hazelcast (ASL 2.0)
> - JUnit (CPL)
> - MigLayout (BSD)
> - Smart and Simple Web Crawler (ASL 2.0)
>
> Dependencies on libraries with non AL-compatible licenses:
> - JFreeChart (LGPL) - runtime dependency only
> - SaverBeans SDK (LGPL) - in the process of being removed
> - JAligner (GPL)
> - Izpack (GPL)
> - NSIS (zlib/libpng, bzip2, CPL) - in the process of being removed
>
> == Cryptography ==
> JPPF does not have any cryptographic component. However, the
> distribution includes a sample that shows encryption/decryption of data
> as a demonstration of one of its features. The sample is delivered with
> full source code and can be found at:
> http://www.jppf.org/wiki/index.php?title=Extending_and_Customizing_JPPF#Transforming_and_encrypting_networked_data
>
> == Required Resources ==
> Mailing lists
>
> * jppf-private (with moderated subscriptions)
> * jppf-dev
> * jppf-commits
> * jppf-user
>
> Subversion Repository
>
> * https://svn.apache.org/repos/asf/incubator/jppf
>
> Issue Tracking
>
> * JIRA (JPPF)
>
> Others
>
> * Web site: Confluence (JPPF)
>
> == Initial Committers ==
> * Laurent Cohen (laurent.cohen at jppf.org)
> * John Channing (john.channing at gmail.com)
>
> == Affiliations ==
> None of the initial committers are paid by their employer, nor do they
> represent their employer in any activity related to JPPF.
>
> == Sponsors ==
> Champion
>
> * Emmanuel Lecharny (elecharny at apache dot org)
>
> Nominated Mentors
>
> We are currently looking for mentors within the community.
>
> Sponsoring Entity
>
> * Apache Incubator
>
>
> --
> Regards,
> Cordialement,
> Emmanuel Lécharny
> www.nextury.com
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
> For additional commands, e-mail: general-help@incubator.apache.org
Jeremias Maerki
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: [Proposal] JPPF : a parallel processing framework for Java
Posted by Emmanuel Lcharny <el...@gmail.com>.
Grant Ingersoll a écrit :
> On Jan 12, 2010, at 10:12 AM, Emmanuel LŽcharny wrote:
>
>
>> Grant Ingersoll a écrit :
>>
>>> Seems like this might fit nicely with Hadoop. Has anyone approached their PMC about sponsoring?
>>>
>>>
>> No, not yet, but that's clearly an option. At least, a better fit than MINA, IMO.
>>
>> Let's do that.
>>
I posted a mail to private@hadoop.apache.org.
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: [Proposal] JPPF : a parallel processing framework for Java
Posted by "Alan D. Cabrera" <li...@toolazydogs.com>.
On Jan 12, 2010, at 11:06 AM, Emmanuel Lcharny wrote:
> Alan D. Cabrera a écrit :
>>
>> On Jan 12, 2010, at 7:27 AM, Grant Ingersoll wrote:
>>
>>>
>>> On Jan 12, 2010, at 10:12 AM, Emmanuel LŽcharny wrote:
>>>
>>>> Grant Ingersoll a écrit :
>>>>> Seems like this might fit nicely with Hadoop. Has anyone
>>>>> approached their PMC about sponsoring?
>>>>>
>>>> No, not yet, but that's clearly an option. At least, a better fit
>>>> than MINA, IMO.
>>>>
>>>> Let's do that.
>>>
>>> Yeah, Hadoop isn't just about Map-Reduce.
>>
>> Just curious, if there's no Hadoop tech in the project then why
>> have it sponsored by Hadoop?
> Well, probably because it would be the place where we can find some
> people interested in participating in the project, increasing the
> community around it. Or maybe merging with Hadoop if there are many
> parts that can be shared, or becoming a subproject if it's
> substancially different in some areas, but have parts in common.
>
> At the end of the day, if Hadoop peeps think that the project is
> orthogonal to what they do, at least they can say so.
Makes sense to me. Thanks.
Regards,
Alan
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: [Proposal] JPPF : a parallel processing framework for Java
Posted by Emmanuel Lcharny <el...@gmail.com>.
Alan D. Cabrera a écrit :
>
> On Jan 12, 2010, at 7:27 AM, Grant Ingersoll wrote:
>
>>
>> On Jan 12, 2010, at 10:12 AM, Emmanuel LŽcharny wrote:
>>
>>> Grant Ingersoll a écrit :
>>>> Seems like this might fit nicely with Hadoop. Has anyone
>>>> approached their PMC about sponsoring?
>>>>
>>> No, not yet, but that's clearly an option. At least, a better fit
>>> than MINA, IMO.
>>>
>>> Let's do that.
>>
>> Yeah, Hadoop isn't just about Map-Reduce.
>
> Just curious, if there's no Hadoop tech in the project then why have
> it sponsored by Hadoop?
Well, probably because it would be the place where we can find some
people interested in participating in the project, increasing the
community around it. Or maybe merging with Hadoop if there are many
parts that can be shared, or becoming a subproject if it's substancially
different in some areas, but have parts in common.
At the end of the day, if Hadoop peeps think that the project is
orthogonal to what they do, at least they can say so.
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: [Proposal] JPPF : a parallel processing framework for Java
Posted by Otis Gospodnetic <ot...@yahoo.com>.
Was my thinking, too. How long before dc.apache.org (or some variant of it) is formed?
Otis
--
Sematext -- http://sematext.com/ -- Solr - Lucene - Nutch
----- Original Message ----
> From: Grant Ingersoll <gs...@apache.org>
> To: general@incubator.apache.org
> Sent: Tue, January 12, 2010 3:50:52 PM
> Subject: Re: [Proposal] JPPF : a parallel processing framework for Java
>
>
> On Jan 12, 2010, at 1:47 PM, Alan D. Cabrera wrote:
>
> >
> > On Jan 12, 2010, at 7:27 AM, Grant Ingersoll wrote:
> >
> >>
> >> On Jan 12, 2010, at 10:12 AM, Emmanuel LŽcharny wrote:
> >>
> >>> Grant Ingersoll a écrit :
> >>>> Seems like this might fit nicely with Hadoop. Has anyone approached their
> PMC about sponsoring?
> >>>>
> >>> No, not yet, but that's clearly an option. At least, a better fit than MINA,
> IMO.
> >>>
> >>> Let's do that.
> >>
> >> Yeah, Hadoop isn't just about Map-Reduce.
> >
> > Just curious, if there's no Hadoop tech in the project then why have it
> sponsored by Hadoop?
> >
>
> I'd add, I think most in Hadoop land view Hadoop as one of the primary places
> for large scale distributed computing at Apache. Map Reduce is one approach and
> it does not fit all situations, so I think you'll see other things arise there,
> possibly JPPF.
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
> For additional commands, e-mail: general-help@incubator.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: [Proposal] JPPF : a parallel processing framework for Java
Posted by Grant Ingersoll <gs...@apache.org>.
On Jan 12, 2010, at 1:47 PM, Alan D. Cabrera wrote:
>
> On Jan 12, 2010, at 7:27 AM, Grant Ingersoll wrote:
>
>>
>> On Jan 12, 2010, at 10:12 AM, Emmanuel LŽcharny wrote:
>>
>>> Grant Ingersoll a écrit :
>>>> Seems like this might fit nicely with Hadoop. Has anyone approached their PMC about sponsoring?
>>>>
>>> No, not yet, but that's clearly an option. At least, a better fit than MINA, IMO.
>>>
>>> Let's do that.
>>
>> Yeah, Hadoop isn't just about Map-Reduce.
>
> Just curious, if there's no Hadoop tech in the project then why have it sponsored by Hadoop?
>
I'd add, I think most in Hadoop land view Hadoop as one of the primary places for large scale distributed computing at Apache. Map Reduce is one approach and it does not fit all situations, so I think you'll see other things arise there, possibly JPPF.
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: [Proposal] JPPF : a parallel processing framework for Java
Posted by "Alan D. Cabrera" <li...@toolazydogs.com>.
On Jan 12, 2010, at 7:27 AM, Grant Ingersoll wrote:
>
> On Jan 12, 2010, at 10:12 AM, Emmanuel LŽcharny wrote:
>
>> Grant Ingersoll a écrit :
>>> Seems like this might fit nicely with Hadoop. Has anyone
>>> approached their PMC about sponsoring?
>>>
>> No, not yet, but that's clearly an option. At least, a better fit
>> than MINA, IMO.
>>
>> Let's do that.
>
> Yeah, Hadoop isn't just about Map-Reduce.
Just curious, if there's no Hadoop tech in the project then why have
it sponsored by Hadoop?
Regards,
Alan
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: [Proposal] JPPF : a parallel processing framework for Java
Posted by Grant Ingersoll <gs...@apache.org>.
On Jan 12, 2010, at 10:12 AM, Emmanuel Lcharny wrote:
> Grant Ingersoll a écrit :
>> Seems like this might fit nicely with Hadoop. Has anyone approached their PMC about sponsoring?
>>
> No, not yet, but that's clearly an option. At least, a better fit than MINA, IMO.
>
> Let's do that.
Yeah, Hadoop isn't just about Map-Reduce.
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: [Proposal] JPPF : a parallel processing framework for Java
Posted by Emmanuel Lcharny <el...@gmail.com>.
Grant Ingersoll a écrit :
> Seems like this might fit nicely with Hadoop. Has anyone approached their PMC about sponsoring?
>
No, not yet, but that's clearly an option. At least, a better fit than
MINA, IMO.
Let's do that.
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: [Proposal] JPPF : a parallel processing framework for Java
Posted by Grant Ingersoll <gs...@apache.org>.
Seems like this might fit nicely with Hadoop. Has anyone approached their PMC about sponsoring?
On Jan 12, 2010, at 7:24 AM, Emmanuel Lecharny wrote:
> Hi guys,
>
> On behalf of the JPPF community, I'd like to bring the following
> proposal for discussion within the Incubator. I had a meeting
> with Laurent last week when he investigated about using MINA into
> JPPF, and I suggested that this project could be a good fit for
> the Incubator.
>
> We will need mentors for this proposal
>
> WDYT ?
>
> Thanks!
>
> http://wiki.apache.org/incubator/JppfProposal
>
> --- Current Wiki Text below ---
> = JPPF : a parallel processing framework for Java =
> === Abstract ===
> JPPF enables applications with large processing power requirements to be
> run on any number of computers, in order to dramatically reduce their
> processing time. This is done by splitting an application into smaller
> parts that can be executed simultaneously on different machines.
>
> === Proposal ===
> JPPF aims at facilitating the parallelization of computation-intensive
> applications, with a focus on ease of use, performance and reliability.
>
> To achieve these goals, the framework comes with a number of outstanding
> features:
>
> - Comprehensive, easy to use APIs: passing from a single-threaded
> application model to a grid-based parallel model can be a daunting task.
> JPPF facilitates this work by providing developers with a set of APIs
> that are simple, can be learned quickly and require a minimal or no
> modification to the existing code.
>
> - No configuration usage: in most environments, JPPF can be deployed
> without any additional configuration burden. Nodes and application
> clients will automatically dicover the servers on the network. The
> server will automatically adapt to workload changes and optimize the
> throughput. Required code and libraries will be automatically deployed
> where they are needed.
>
> - Dynamic grid scaling and self-repair: the JPPF grid is fault-tolerant,
> meaning that the failure of a node, or even a server, does not
> compromise the jobs currently executing or scheduled. In most cases, the
> performance degradation will be barely noticeable, as JPPF automatically
> adapts to topology and workload changes. Furthermore, nodes and servers
> can be dynamically started and will be automatically recognized,
> allowing JPPF to function in "crunch mode". In addition to this, JPPF
> components benefit from automatic recovery functionalities.
>
> - Job-level SLA: each job submitted to the JPPF grid runs within limits
> defined by its own SLA (service level agreement). This allows to specify
> the characteristics (i.e. available memory, processors, disk space,
> operating systems, etc.) of the nodes a job can run on, as well as how
> many nodes it can run on. As many functionalities in JPPF, this one can
> be dynamically adjusted, manually or automatically.
>
> - Advanced Management and monitoring: full-fledged management and
> monitoring features are provided out of the box: server and nodes status
> monitoring, detailed statistics and events, remote administration,
> job-level real-time monitoring and management, charts, cpu utilization
> (for billing). These functionalities are available via a graphical user
> interface as well as from the JPPF APIs.
>
> - Integration with leading application and web servers: by complying
> with the Java Connector Architecture 1.5 specification, JPPF integrates
> seamlessly with and completes the offering of leading J2EE application
> servers: Apache Geronimo, JBoss, Glassfish, IBM Websphere, Oracle
> Weblogic, Oracle OC4J. JPPF also integrates with GigaSpaces eXtreme
> Application Platform and Apache Tomcat web server
>
> - Extensibility: JPPF offers a number of hooks and extension points that
> allow users to extend the framework and adapt it to their own
> requirements and needs. Such extension mechanisms are available for
> custom management and monitoring MBeans, startup classes for grod
> components, network data transformation/encryption, additional
> load-balancing algorithms, alternate object serialization mechanisms.
>
> === Background ===
> JPPF is a project being actively developed at SourceForge. It was
> created to address a class of problems called "embarassingly parallel",
> which groups computational problems that can be decomposed into many
> smaller sub-problems, that are independant from each other and that can
> thus be executed in parallel.
>
> === Rationale ===
> Given these last years' emergence of technologies that make commodity
> hardware, virtualization and cloud computing available to a fast-growing
> computing ecosystem, the project answers the need to execute
> applications ever faster, with a low entry cost, while at the same time
> preserving historical technological investments.
>
> == Current Status ==
> === Meritocracy ===
> MWe acknowledge that a meritocratic governance is the only way for the
> project to grow and expand, in the spirit of open source and the ASF. It
> will benefit the project, its communities, the ASF, as well as the outer
> ecosystems.
>
> === Community ===
> JPPF already has a relatively small, but steadily growing community of
> users. Given the applicability of the project to numerous industries and
> technological areas such as scientific research, finance,
> graphical/video rendering, telecoms, data mining, etc.., we are
> confident that there is a very large growth potential for developers and
> users communities around JPPF.
>
> === Core Developers ===
> JPPF was founded in April 2005 by Laurent Cohen (laurent.cohen at
> jppf.org), who is currently the only active developer and code committer.
> The other active contributor is John Channing (john.channing at
> gmail.com) who provides advice and peer review on the project's
> architecture, design, requirements and promotion.
> Other project members and former project contributors include:
> - Domingos Creado (dcreado at users.sourceforge.net)
> - Mahendra Kutare (mahendra.kutare at gmail.com)
> - Guy Korland (Guy at gigaspaces.com)
> - Fredreric Barachant (pepe-barachant at users.sourceforge.net)
> - Peter Becker (nc-heuelhe at netcologne.de)
> - Wolfgang Wagner (wolfgang.wagner at iname.com)
> - Jay Yusko (jay.yusko at gensolco.com)
>
> === Alignment ===
> The project is currently undergoing a refactoring of its network
> communication infrastructure, relying essentially on Mina, in order to
> provide a greater maintainability of the code, as well as extend its
> scalability and provide new functionalities such as truly secure
> communications and integration of other distributed computing models
> (e.g. P2P, map/reduce).
> JPPF also offers connectors for Geronimo and Tomcat.
>
> == Known Risks ==
> === Orphaned products ===
> This is the main and most obvious risk of this project. The knowledge of
> the code is owned by a single developer, Laurent Cohen, who has
> committed about 98% of the existing code in the current repository. We
> are well aware of this problem, and it is our hope and challenge that
> integrating the Apache community will inspire the growth of a strong
> developers community around JPPF.
>
> === Inexperience with Open Source ===
> JPPF was created as an Open Source project in April 2005, and has
> remained so since.
>
> === Homogenous Developers ===
> A single active developer implies homogeneity.
>
> === Reliance on Salaried Developers ===
> ANo one is currently paid to work on JPPF.
>
> === Relationships with Other Apache Products ===
> JPPF is currently refactoring its network communication infrastructure
> with the help of Mina (which is how we came to find a sponsor).
> The project's logging relies entirely on commons-logging and Log4j.
> The project also provides connectors for Geronimo and Tomcat.
> Builds are done with Ant, and we are currently looking at switching to
> Maven-based builds.
> We foresee that JPPF could be used by many other Apache projects to
> speed-up the execution of unit tests by running them in parallel, and to
> serve as a basis for running distributed charge and load tests.
> One of our samples uses Lucene to demonstrate a simple use of a
> distributed framework for web search, crawling and indexing
>
> === A Excessive Fascination with the Apache Brand ===
> Our main hope in joining the Apache community is that it will help build
> a strong developers and committers community around the project, and
> remove its reliance on a single developer. We aim at achieving this
> while attracting new users, in conformance with the Apache spirit and
> policies.
>
> == Documentation ==
> Information on JPPF can be found at:
> http://www.jppf.org (main site, documentation and user forums)
> http://sourceforge.net/projects/jppf-project (code repository, bug
> tracker, features tracker)
>
> == Initial Source ==
> OThe entire code for JPPF is held in a SourceForge CVS repository, and
> has been so since the project's inception in April 2005. This entire
> code base will be donated to Apache.
> The source code has been licensed under the ASL 2.0 since August 2007.
> Before that it was LGPL.
> All artifacts in the trunk and existing branches (1 branch) are now
> licensed under the ASL 2.0, with corresponding header when applicable.
>
> == External Dependencies ==
> Other than the Apache products used in JPPF, the project currently
> depends on the following ASL 2.0 compatible products/licenses:
> - Rhino (MPL)
> - JGoodies Looks (BSD)
> - Groovy (ASL 2.0)
> - Hazelcast (ASL 2.0)
> - JUnit (CPL)
> - MigLayout (BSD)
> - Smart and Simple Web Crawler (ASL 2.0)
>
> Dependencies on libraries with non AL-compatible licenses:
> - JFreeChart (LGPL) - runtime dependency only
> - SaverBeans SDK (LGPL) - in the process of being removed
> - JAligner (GPL)
> - Izpack (GPL)
> - NSIS (zlib/libpng, bzip2, CPL) - in the process of being removed
>
> == Cryptography ==
> JPPF does not have any cryptographic component. However, the
> distribution includes a sample that shows encryption/decryption of data
> as a demonstration of one of its features. The sample is delivered with
> full source code and can be found at:
> http://www.jppf.org/wiki/index.php?title=Extending_and_Customizing_JPPF#Transforming_and_encrypting_networked_data
>
> == Required Resources ==
> Mailing lists
>
> * jppf-private (with moderated subscriptions)
> * jppf-dev
> * jppf-commits
> * jppf-user
>
> Subversion Repository
>
> * https://svn.apache.org/repos/asf/incubator/jppf
>
> Issue Tracking
>
> * JIRA (JPPF)
>
> Others
>
> * Web site: Confluence (JPPF)
>
> == Initial Committers ==
> * Laurent Cohen (laurent.cohen at jppf.org)
> * John Channing (john.channing at gmail.com)
>
> == Affiliations ==
> None of the initial committers are paid by their employer, nor do they
> represent their employer in any activity related to JPPF.
>
> == Sponsors ==
> Champion
>
> * Emmanuel Lecharny (elecharny at apache dot org)
>
> Nominated Mentors
>
> We are currently looking for mentors within the community.
>
> Sponsoring Entity
>
> * Apache Incubator
>
>
> --
> Regards,
> Cordialement,
> Emmanuel Lécharny
> www.nextury.com
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
> For additional commands, e-mail: general-help@incubator.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: [Proposal] JPPF : a parallel processing framework for Java
Posted by "Alan D. Cabrera" <li...@toolazydogs.com>.
Happy to help a a mentor.
Regards,
Alan
On Jan 12, 2010, at 4:24 AM, Emmanuel Lecharny wrote:
> Hi guys,
>
> On behalf of the JPPF community, I'd like to bring the following
> proposal for discussion within the Incubator. I had a meeting
> with Laurent last week when he investigated about using MINA into
> JPPF, and I suggested that this project could be a good fit for
> the Incubator.
>
> We will need mentors for this proposal
>
> WDYT ?
>
> Thanks!
>
> http://wiki.apache.org/incubator/JppfProposal
>
> --- Current Wiki Text below ---
> = JPPF : a parallel processing framework for Java =
> === Abstract ===
> JPPF enables applications with large processing power requirements
> to be
> run on any number of computers, in order to dramatically reduce their
> processing time. This is done by splitting an application into smaller
> parts that can be executed simultaneously on different machines.
>
> === Proposal ===
> JPPF aims at facilitating the parallelization of computation-intensive
> applications, with a focus on ease of use, performance and
> reliability.
>
> To achieve these goals, the framework comes with a number of
> outstanding
> features:
>
> - Comprehensive, easy to use APIs: passing from a single-threaded
> application model to a grid-based parallel model can be a daunting
> task.
> JPPF facilitates this work by providing developers with a set of APIs
> that are simple, can be learned quickly and require a minimal or no
> modification to the existing code.
>
> - No configuration usage: in most environments, JPPF can be deployed
> without any additional configuration burden. Nodes and application
> clients will automatically dicover the servers on the network. The
> server will automatically adapt to workload changes and optimize the
> throughput. Required code and libraries will be automatically deployed
> where they are needed.
>
> - Dynamic grid scaling and self-repair: the JPPF grid is fault-
> tolerant,
> meaning that the failure of a node, or even a server, does not
> compromise the jobs currently executing or scheduled. In most cases,
> the
> performance degradation will be barely noticeable, as JPPF
> automatically
> adapts to topology and workload changes. Furthermore, nodes and
> servers
> can be dynamically started and will be automatically recognized,
> allowing JPPF to function in "crunch mode". In addition to this, JPPF
> components benefit from automatic recovery functionalities.
>
> - Job-level SLA: each job submitted to the JPPF grid runs within
> limits
> defined by its own SLA (service level agreement). This allows to
> specify
> the characteristics (i.e. available memory, processors, disk space,
> operating systems, etc.) of the nodes a job can run on, as well as how
> many nodes it can run on. As many functionalities in JPPF, this one
> can
> be dynamically adjusted, manually or automatically.
>
> - Advanced Management and monitoring: full-fledged management and
> monitoring features are provided out of the box: server and nodes
> status
> monitoring, detailed statistics and events, remote administration,
> job-level real-time monitoring and management, charts, cpu utilization
> (for billing). These functionalities are available via a graphical
> user
> interface as well as from the JPPF APIs.
>
> - Integration with leading application and web servers: by complying
> with the Java Connector Architecture 1.5 specification, JPPF
> integrates
> seamlessly with and completes the offering of leading J2EE application
> servers: Apache Geronimo, JBoss, Glassfish, IBM Websphere, Oracle
> Weblogic, Oracle OC4J. JPPF also integrates with GigaSpaces eXtreme
> Application Platform and Apache Tomcat web server
>
> - Extensibility: JPPF offers a number of hooks and extension points
> that
> allow users to extend the framework and adapt it to their own
> requirements and needs. Such extension mechanisms are available for
> custom management and monitoring MBeans, startup classes for grod
> components, network data transformation/encryption, additional
> load-balancing algorithms, alternate object serialization mechanisms.
>
> === Background ===
> JPPF is a project being actively developed at SourceForge. It was
> created to address a class of problems called "embarassingly
> parallel",
> which groups computational problems that can be decomposed into many
> smaller sub-problems, that are independant from each other and that
> can
> thus be executed in parallel.
>
> === Rationale ===
> Given these last years' emergence of technologies that make commodity
> hardware, virtualization and cloud computing available to a fast-
> growing
> computing ecosystem, the project answers the need to execute
> applications ever faster, with a low entry cost, while at the same
> time
> preserving historical technological investments.
>
> == Current Status ==
> === Meritocracy ===
> MWe acknowledge that a meritocratic governance is the only way for the
> project to grow and expand, in the spirit of open source and the
> ASF. It
> will benefit the project, its communities, the ASF, as well as the
> outer
> ecosystems.
>
> === Community ===
> JPPF already has a relatively small, but steadily growing community of
> users. Given the applicability of the project to numerous industries
> and
> technological areas such as scientific research, finance,
> graphical/video rendering, telecoms, data mining, etc.., we are
> confident that there is a very large growth potential for developers
> and
> users communities around JPPF.
>
> === Core Developers ===
> JPPF was founded in April 2005 by Laurent Cohen (laurent.cohen at
> jppf.org), who is currently the only active developer and code
> committer.
> The other active contributor is John Channing (john.channing at
> gmail.com) who provides advice and peer review on the project's
> architecture, design, requirements and promotion.
> Other project members and former project contributors include:
> - Domingos Creado (dcreado at users.sourceforge.net)
> - Mahendra Kutare (mahendra.kutare at gmail.com)
> - Guy Korland (Guy at gigaspaces.com)
> - Fredreric Barachant (pepe-barachant at users.sourceforge.net)
> - Peter Becker (nc-heuelhe at netcologne.de)
> - Wolfgang Wagner (wolfgang.wagner at iname.com)
> - Jay Yusko (jay.yusko at gensolco.com)
>
> === Alignment ===
> The project is currently undergoing a refactoring of its network
> communication infrastructure, relying essentially on Mina, in order to
> provide a greater maintainability of the code, as well as extend its
> scalability and provide new functionalities such as truly secure
> communications and integration of other distributed computing models
> (e.g. P2P, map/reduce).
> JPPF also offers connectors for Geronimo and Tomcat.
>
> == Known Risks ==
> === Orphaned products ===
> This is the main and most obvious risk of this project. The
> knowledge of
> the code is owned by a single developer, Laurent Cohen, who has
> committed about 98% of the existing code in the current repository. We
> are well aware of this problem, and it is our hope and challenge that
> integrating the Apache community will inspire the growth of a strong
> developers community around JPPF.
>
> === Inexperience with Open Source ===
> JPPF was created as an Open Source project in April 2005, and has
> remained so since.
>
> === Homogenous Developers ===
> A single active developer implies homogeneity.
>
> === Reliance on Salaried Developers ===
> ANo one is currently paid to work on JPPF.
>
> === Relationships with Other Apache Products ===
> JPPF is currently refactoring its network communication infrastructure
> with the help of Mina (which is how we came to find a sponsor).
> The project's logging relies entirely on commons-logging and Log4j.
> The project also provides connectors for Geronimo and Tomcat.
> Builds are done with Ant, and we are currently looking at switching to
> Maven-based builds.
> We foresee that JPPF could be used by many other Apache projects to
> speed-up the execution of unit tests by running them in parallel,
> and to
> serve as a basis for running distributed charge and load tests.
> One of our samples uses Lucene to demonstrate a simple use of a
> distributed framework for web search, crawling and indexing
>
> === A Excessive Fascination with the Apache Brand ===
> Our main hope in joining the Apache community is that it will help
> build
> a strong developers and committers community around the project, and
> remove its reliance on a single developer. We aim at achieving this
> while attracting new users, in conformance with the Apache spirit and
> policies.
>
> == Documentation ==
> Information on JPPF can be found at:
> http://www.jppf.org (main site, documentation and user forums)
> http://sourceforge.net/projects/jppf-project (code repository, bug
> tracker, features tracker)
>
> == Initial Source ==
> OThe entire code for JPPF is held in a SourceForge CVS repository, and
> has been so since the project's inception in April 2005. This entire
> code base will be donated to Apache.
> The source code has been licensed under the ASL 2.0 since August 2007.
> Before that it was LGPL.
> All artifacts in the trunk and existing branches (1 branch) are now
> licensed under the ASL 2.0, with corresponding header when applicable.
>
> == External Dependencies ==
> Other than the Apache products used in JPPF, the project currently
> depends on the following ASL 2.0 compatible products/licenses:
> - Rhino (MPL)
> - JGoodies Looks (BSD)
> - Groovy (ASL 2.0)
> - Hazelcast (ASL 2.0)
> - JUnit (CPL)
> - MigLayout (BSD)
> - Smart and Simple Web Crawler (ASL 2.0)
>
> Dependencies on libraries with non AL-compatible licenses:
> - JFreeChart (LGPL) - runtime dependency only
> - SaverBeans SDK (LGPL) - in the process of being removed
> - JAligner (GPL)
> - Izpack (GPL)
> - NSIS (zlib/libpng, bzip2, CPL) - in the process of being removed
>
> == Cryptography ==
> JPPF does not have any cryptographic component. However, the
> distribution includes a sample that shows encryption/decryption of
> data
> as a demonstration of one of its features. The sample is delivered
> with
> full source code and can be found at:
> http://www.jppf.org/wiki/index.php?title=Extending_and_Customizing_JPPF#Transforming_and_encrypting_networked_data
>
> == Required Resources ==
> Mailing lists
>
> * jppf-private (with moderated subscriptions)
> * jppf-dev
> * jppf-commits
> * jppf-user
>
> Subversion Repository
>
> * https://svn.apache.org/repos/asf/incubator/jppf
>
> Issue Tracking
>
> * JIRA (JPPF)
>
> Others
>
> * Web site: Confluence (JPPF)
>
> == Initial Committers ==
> * Laurent Cohen (laurent.cohen at jppf.org)
> * John Channing (john.channing at gmail.com)
>
> == Affiliations ==
> None of the initial committers are paid by their employer, nor do they
> represent their employer in any activity related to JPPF.
>
> == Sponsors ==
> Champion
>
> * Emmanuel Lecharny (elecharny at apache dot org)
>
> Nominated Mentors
>
> We are currently looking for mentors within the community.
>
> Sponsoring Entity
>
> * Apache Incubator
>
>
> --
> Regards,
> Cordialement,
> Emmanuel Lécharny
> www.nextury.com
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
> For additional commands, e-mail: general-help@incubator.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org