incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Eli Collins <...@cloudera.com>
Subject Re: [PROPOSAL] Oozie for the Apache Incubator
Date Fri, 24 Jun 2011 23:25:57 GMT
+1

Spectacular!

On Fri, Jun 24, 2011 at 12:46 PM, Mohammad Islam <mislam77@yahoo.com> wrote:
> Hi,
>
> I would like to propose Oozie to be an Apache Incubator project.
> Oozie is a server-based workflow scheduling and coordination system to manage
> data processing jobs for Apache Hadoop.
>
>
> Here's a link to the proposal in the Incubator wiki
> http://wiki.apache.org/incubator/OozieProposal
>
>
> I've also pasted the initial contents below.
>
> Regards,
>
> Mohammad Islam
>
>
> Start of Oozie Proposal
>
> Abstract
> Oozie is a server-based workflow scheduling and coordination system to manage
> data processing jobs for Apache HadoopTM.
>
> Proposal
> Oozie is an  extensible, scalable and reliable system to define, manage,
> schedule,  and execute complex Hadoop workloads via web services. More
> specifically, this includes:
>
>        * XML-based declarative framework to specify a job or a complex workflow of
> dependent jobs.
>
>        * Support different types of job such as Hadoop Map-Reduce, Pipe, Streaming,
> Pig, Hive and custom java applications.
>
>        * Workflow scheduling based on frequency and/or data availability.
>        * Monitoring capability, automatic retry and failure handing of jobs.
>        * Extensible and pluggable architecture to allow arbitrary grid programming
> paradigms.
>
>        * Authentication, authorization, and capacity-aware load throttling to allow
> multi-tenant software as a service.
>
> Background
> Most data  processing applications require multiple jobs to achieve their goals,
> with inherent dependencies among the jobs. A dependency could be  sequential,
> where one job can only start after another job has finished.  Or it could be
> conditional, where the execution of a job depends on the  return value or status
> of another job. In other cases, parallel  execution of multiple jobs may be
> permitted – or desired – to exploit  the massive pool of compute nodes provided
> by Hadoop.
>
> These  job dependencies are often expressed as a Directed Acyclic Graph, also
> called a workflow. A node in the workflow is typically a job (a  computation on
> the grid) or another type of action such as an eMail  notification. Computations
> can be expressed in map/reduce, Pig, Hive or  any other programming paradigm
> available on the grid. Edges of the graph  represent transitions from one node
> to the next, as the execution of a  workflow proceeds.
>
> Describing  a workflow in a declarative way has the advantage of decoupling job
> dependencies and execution control from application logic. Furthermore,  the
> workflow is modularized into jobs that can be reused within the same  workflow
> or across different workflows. Execution of the workflow is  then driven by a
> runtime system without understanding the application  logic of the jobs. This
> runtime system specializes in reliable and  predictable execution: It can retry
> actions that have failed or invoke a  cleanup action after termination of the
> workflow; it can monitor  progress, success, or failure of a workflow, and send
> appropriate alerts  to an administrator. The application developer is relieved
> from  implementing these generic procedures.
>
> Furthermore,  some applications or workflows need to run in periodic intervals
> or  when dependent data is available. For example, a workflow could be  executed
> every day as soon as output data from the previous 24 instances  of another,
> hourly workflow is available. The workflow coordinator  provides such scheduling
> features, along with prioritization, load  balancing and throttling to optimize
> utilization of resources in the  cluster. This makes it easier to maintain,
> control, and coordinate  complex data applications.
>
> Nearly  three years ago, a team of Yahoo! developers addressed these critical
> requirements for Hadoop-based data processing systems by developing a  new
> workflow management and scheduling system called Oozie. While it was  initially
> developed as a Yahoo!-internal project, it was designed and  implemented with
> the intention of open-sourcing. Oozie was released as a GitHub project in early
> 2010. Oozie is used in production within Yahoo and  since it has been
> open-sourced it has been gaining adoption with  external developers
>
> Rationale
> Commonly,  applications that run on Hadoop require multiple Hadoop jobs in order
> to  obtain the desired results. Furthermore, these Hadoop jobs are commonly  a
> combination of Java map-reduce jobs, Streaming map-reduce jobs, Pipes
> map-reduce jobs, Pig jobs, Hive jobs, HDFS operations, Java programs  and shell
> scripts.
>
> Because  of this, developers find themselves writing ad-hoc glue programs to
> combine these Hadoop jobs. These ad-hoc programs are difficult to  schedule,
> manage, monitor and recover.
>
> Workflow  management and scheduling is an essential feature for large-scale data
> processing applications. Such applications could write the customized  solution
> that would require separate development, operational, and  maintenance overhead.
> Since it is a prevalent use-case for data  processing, the application developer
> would surely prefer a generalized  solution with little or no such overhead.
> Oozie addresses the challenge  by providing an execution framework to flexibly
> specify the job  dependency, data dependency, and time dependency. In addition,
> Oozie  provides a multi-tenant-based centralized service and the opportunity to
> optimize load and utilization while respecting SLAs.
>
> Oozie  is built on Apache Hadoop to schedule jobs related to various Apache
> projects such as Hadoop, Pig, and Hive. As an Apache Open source  project, Oozie
> is expected to attract the larger and more diversified  community that currently
> uses such Apache sponsored projects.  Additionally, users of the Hadoop
> ecosystem can influence Oozie’s  roadmap, and contribute to it. Likewise, Oozie,
> as part of the Apache  Hadoop ecosystem, will be a great benefit to the current
> Hadoop/Pig/Hive/HBase/HCatalog community.
>
> Current Status
> Meritocracy
> Oozie  currently is a github-based open sourced project where developers from
> multiple companies are contributing to the project. Our intent with this
> incubator proposal is to further extend this diverse developer  community around
> Oozie following the Apache meritocracy model. We plan  to continue to provide
> adequate support to new developers and to quickly  recruit those who make solid
> contributions to committer status. In  addition, Oozie will expect, accept, and
> work to attract contributions  from amateurs as well.
>
> Community
> While an  efficient workflow management and scheduling system is critical for
> large companies with huge data processing in multi-tenant clusters, it  is
> equally necessary for any non-trivial deployment. Different companies  are
> currently using Oozie as a workflow scheduler for Hadoop-based data  processing.
> At Yahoo! it is being used extensively in production  clusters to process
> thousand of jobs. Like the Oozie user community, the  Oozie developer community
> is also very strong. Developers from Yahoo!  provided the initial code base, and
> they are still the most active  contributors. In late 2010, developers from
> Cloudera also started  contributing, and currently other companies (e.g., IBM)
> are beginning to  participate.
>
> We currently use JIRA for issue tracking, github for code hosting and Yahoo!
> Groups for developer and user communications.
>
> Core Developers
> Oozie is  currently being designed and developed by four engineers from Yahoo! –
> Mohammad Islam, Angelo Huang, Mayank Bansal, and Andreas Neumann. In  addition,
> many outside contributors are actively contributing in design  and development.
> Among them, Alejandro Abdelnur from Cloudera and Chao  Wang from IBM are very
> important contributors. All of these core  developers have deep expertise in
> Hadoop and the Hadoop Ecosystem in  general.
>
> Alignment
> The ASF is a  natural host for Oozie given that it is already the home of
> Hadoop,  Pig, Hive, and other emerging cloud software projects. Oozie was
> designed to support Hadoop from the beginning in order to solve data  processing
> challenges in Hadoop clusters. Oozie complements the existing  Apache cloud
> computing projects by providing a flexible framework for  managing complex data
> processing tasks.
>
> Known Risks
> Orphaned Products
> The core  developers plan to work full time on the project. There is very little
> risk of Oozie getting orphaned since large companies like Yahoo! are
> extensively using it on their production Hadoop clusters. For example,  there
> are nearly 400 Yahoo! internal Oozie users and thousands of jobs  are processed
> hourly through Oozie in production. In addition, there are  nearly 400 active
> users (including Yahoo! internal and external) in the  email community where
> nearly 15 emails are exchanged per day.  Furthermore, there were more than 1500
> downloads of the Oozie binary in  the last eight months from the github site and
> a large number of  downloads were conducted by other companies such as Cloudera.
> Oozie has  three major releases and more than 15 patch releases in the last
> couple  of years which further demonstrates Oozie as a very active project. We
> plan to extend and diversify this community further through Apache.
>
> Inexperience with Open Source
> The core  developers are all active users and followers of open source. They are
> already committers and contributors to the Oozie Github project. In  addition,
> they are very familiar with Apache principals and philosophy  for community
> driven software development.
>
> Homogeneous Developers
> The core developers are from Yahoo! as well as from several other corporations,
> including Cloudera and IBM.
>
> Reliance on Salaried Developers
> Currently,  the developers are paid to do work on Oozie. Companies like Yahoo!
> and  Cloudera are invested in Oozie as the solution to the workflow  management
> and scheduling problem in Hadoop clusters, and that is not  likely to change. In
> addition, since workflow management is very  important for most hadoop based
> data processing, non-salaried developers  and researchers from various
> institutes are expected to contribute to  the project.
>
> Relationships with Other Apache Products
> Oozie is  based on Apache Hadoop to manage jobs created by different Apache
> projects such as Hadoop, Pig, and Hive. Users of these products are  extensively
> using Oozie as their workflow scheduler.
>
> An Excessive Fascination with the Apache Brand
> We deeply  respect the reputation of Apache and have had great success with
> other  Apache projects such as Pig and HCatalog. We are motivated to expand and
> increase the adoption and development of Oozie following Apache’s  established
> open source model. We have also given reasons in the  Rationale and Alignment
> sections.
>
> Documentation
> Information about Oozie can be found at http://yahoo.github.com/oozie/. The
> following links provide more information about Oozie in open source:
>
>        * Codebase at GitHub: https://github.com/yahoo/oozie.
>        * JIRA : http://oozie-jira.hadoop.developer.yahoo.net
>        * Continuous Integration (CI) build:
> http://oozie-ci.hadoop.developer.yahoo.net/
>
>        * Yahoo user community: http://tech.groups.yahoo.com/group/Oozie-users/
> Initial Source
> Oozie has been under development since 2009 by a team of engineers at Yahoo!. It
> is currently hosted on GitHub under an Apache license at
> https://github.com/yahoo/oozie.
>
> External Dependencies
> The required  external dependencies are all Apache License or compatible
> licenses.  Following the components with non-Apache licenses are enumerated:
>
>        * HSQLDB License: HSQLDB
>        * JDOM license: JDOM
>        * BSD: Serp
>        * CCDL v1: jaxb-api, ejb, JAF
> NOTE:  With the exception of HSQLDB and JDOM that are directly used by Oozie,
> the other listed components are transitive dependencies of other Apache
> components used by Oozie.
>
> Cryptography
> Oozie supports the Kerberos authentication mechanism to access secured Hadoop
> services.
>
> Required Resources
> Mailing Lists
>        * oozie-private for private PMC discussions (with moderated subscriptions)
>        * oozie-dev
>        * oozie-commits
>        * oozie-user
> Subversion Directory
> https://svn.apache.org/repos/asf/incubator/oozie
> Issue Tracking
> JIRA Oozie (OOZIE)
> Other Resources
> The  existing code already has unit tests, so we would like a Hudson instance
> to run them whenever a new patch is submitted. This can be added after  project
> creation.
>
> Initial Committers
>        * Mohammad K Islam (mislam77 at yahoo  dot com)
>        * Angelo K Huang (angelohuang at gmail dot com)
>        * Mayank Bansal (mabansal at gmail dot com)
>        * Andreas Neumann (neunand at gmail dot com)
>        * Alejandro Abdelnur (tucu00 at gmail dot com)
>        * Chao Wang (brookwc at gmail dot com)
> Affiliations
>        * Mohammad K Islam (Yahoo!)
>        * Angelo Huang (Yahoo!)
>        * Mayank Bansal (Yahoo!)
>        * Andreas Neumann (Yahoo!)
>        * Alejandro Abdelnur (Cloudera)
>        * Chao Wang (IBM)
> Sponsors
> Champion
> Alan Gates
> Nominated Mentors
>        * Owen O'Malley (Incubator PMC member)
>        * Alan Gates (Incubator PMC member)
>        * Christopher Douglas(Incubator PMC member)
>        * Devaraj Das (Hadoop PMC member)
> Sponsoring EntityWe are requesting the Incubator to sponsor this project.

---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org


Mime
View raw message