Apache Hadoop, a top level apache project [1] is a Java software framework that supports data-intensive distributed applications under an Open Source license.
Talend is a the recognized market leader in open source data integration and we achieved a first step of Hadoop technologies (HIVE, HDFS, HIVE, PIG, Scooq...) integration [2].
The goal of this project is to extend the support of Hadoop thanks coding some optimize HIVE SQL programs (HIVE Templates), Pig scripts and benchmark in a grid of servers. You will get some high knowledge of hadoop after this project will be done.
You will work daily with Chinese local leader in Talend Beijing location (80 developers) and report to an English spoken project manager based in France.
[1] http://hadoop.apache.org/
[2] http://cn.talend.com/products-data-integration/talend-integration-suite-mpx.php#feature

As the emergence of cloud computing, millions of corporations and organizations had put their service on cloud. In this topic, we need re-implement Bench4Q tool as a service (BaaS) deploying on the application server (JOnAS). BaaS should support elasticity and multi-tenancy which are promised in cloud computing. The tenants of BaaS are supposed to deliver the address of system under test and descriptions of the stress testing to BaaS, and then receive reports on system under test after testing without concerning about managing and coordinating the testing resource beneath the cloud computing platform.

As the scalability and management limitation of traditional relational database, more and more new coming Web applications is designed for NoSQL stores, such as key-value store, document-style store and so on. The subject consists in migrating the TPC-W online bookstore application to key-value stores. Any other applications for using NoSQL stores are welcome, for example, social networks and online games.

With the development of Internet, there is lots of information on the
Internet. Many applications have been developed to collection data (e.g.
web pages) from the Internet, and try to discover useful information or
knowledge from the collected data. The most popular examples of such kind
of applications are search engines, like Google, Bing, etc. However, in
most cases, the crawlers used in these applications cannot understand the
contents of the web pages, so they just download the whole pages and leave
the parsing work for humans or other applications. And in fact the parsing
work is very tedious. Recently, Semantic Web technologies have been
proposed and adopted in many applications. In Semantic Web, the computers
can understand the contents of the web pages. It will be very useful to
build some kind of semantic-enabled web crawler and parser that can
understand the contents of the web pages. The goal of this topic/project is
to build such kind web crawler and parser.

First integrate Eclipse Nebula TreeMapper in Scarbo's JWT workflow editor. Then allow to specify expressions within workflow variables (first XPath for XML, if possible javascript for JSON or other kind of data) in it when mapping data, and enhance the runtime's DataMappingService so it first evaluates them. Finally provide a demo sample and wiki doc.

Study and prototype Scarbo monitoring using OW2 Jasmine. Expose Scarbo runtime's WorkflowService implementation features as JMX bean(s), and access it within Jasmine using JMX probes. Write rules detecting when there are "too much" (more than a set amount) or "forgotten" (not accessed in the previous 3 months) running instances of a workflow model and propose other interesting ones.
Then allow these rules to be parameterized and configured at design time. For this, write an Eclipse plugin for Scarbo's JWT workflow editor defining a model extension where those parameters can be stored, and enhance said JMX beans to load these values at runtime using the ProcessModelService (from the Transformation's property file output).
Provide a prototype, demonstrating samples and wiki doc.

Build a web conferencing system based on standard HTML5 technologies, especially on WebRTC (still in early development).
The aim is for users to engage a conversation (audio or video) on a specific subject (part of a product structure for instance), his partner should be informed of the context of the discussion. At the end, the system should be able to record the conversation in order to capitalize on it. Optionally, we should be able to integrate a third party tool like Dragon speaking in order to get the verbatim.

The WebLab is an open source (under LGPL 2.1) platform aimed at providing intelligence systems that need to process multimedia data. Thus a system based on WebLab tackle the problem of “unstructured document processing” and in particular in the analysis of documents coming for the Internet. One of its typical application is media monitoring which could serve many different business needs.

In that context, a particular problem is that content found on the web may be protected by specific terms of use and/or copyrights which could prevent (or limit) the use of such content. Finding the legal terms applied on content is thus a very important operational issue and currently only manual assessment allows to find these legal terms.

The objective of this programming contest would be to try to automate the detection of the terms of use and copyright of content found on the web. This work would be organised in 3 steps:
1- learn to use a weblab system to collect and process content from website ;
2- try to implement a specific analysis service that tries to detect if a page contains legal terms. This problem will probably need the manual annotation of documents in order to create a corpus of samples. Then the analysis service will apply different technique (rule based to NLP) in order to tag document ;
3- when a document is not tagged as being “legal terms”, one will needs to detect possible related pages or linked that point to legal terms that could be applied. This will probably include analysis of website map and link graph to apply specific algorithms that allows to deduce which terms apply to what documents.

The WebLab is an open source (under LGPL 2.1) platform aimed at providing intelligence systems that need to process multimedia data. Thus a system based on WebLab tackle the problem of “unstructured document processing” and in particular in the analysis of documents coming for the Internet. One of its typical application is media monitoring which could serve many different business needs.

Navigation and manipulation of multimedia content is one of the key possibility offered by the platform. This project focussed on the presentation of such content. The objective will be to develop specific use interface to present and navigate within video that has been processed using latest web technologies (HTML5 and javascript) and integrating it into a JSR 286 portlet. This portlet will consume event and data from WebLab and will enable the user to exploit the video and possible information that has been extracted.

The WebLab is an open source (under LGPL 2.1) platform aimed at providing intelligence systems that need to process multimedia data. Thus a system based on WebLab tackle the problem of “unstructured document processing” and in particular in the analysis of documents coming for the Internet. One of its typical application is media monitoring which could serve many different business needs.

One of the problem faced while processing information from the web is the large amount of data that is created. The architecture allows to easily distribute processing power and duplicate services. This project will then focus on the processing part and explore the possibility to integrate distributed storage and in particular explore Hadoop technologies and relevant sub projects (HDFS and Cassandra). The goal will be to study the integration of such storage capabilities as WebLab service for the multiple data types encountered: raw data from the Web (text, audio, images, video), XML and RDF triples).

At the moment, the Telosys framework provides a Record Oriented Persistence (“ROP”) system for classical Relational Databases (based on DAO using SQL and JDBC). The Telosys Tools (Eclipse Plugin) can be used to generate all the code of the DAO classes.

The goal of this new project is to provide a persistence system for HBase (the Hadoop NoSQL database). The API must be the same as the existing DAO in order to offer a transparent access to HBase. The persistence implementation must not required specific coding for the user (it must be either based on reflection or generated with the Eclipse plugin). The Hbase API must be used to store, retrieve and delete any kind of Java “Value Object” (VO/DTO). A new query criteria system will have to be designed.

OW2 Petals Entreprise Service Bus is moving to the Cloud to provide the so-famous Cloud Service Bus.
Such service platform needs to be managed from the SaaS layer. We propose to manage Cloud Service Bus nodes and services from a nice and innovative Web application running in the Cloud.

The project consists in integrating a mapping widget in the Eclipse BPEL Designer.
As a reminder, BPEL is a standard used to build web services by assembling other web services.
This mapping widget already exists and is part of the Eclipse Nebula project. It should be reused in the BPEL Designer to define XML mapping between service inputs and service outputs. The mapper will have to deal with simple assignations (the target and the source have the same type), static assignations (assign a fixed value to a BPEL variable) and complex assignations (using XPath expressions and functions).

This project is an opportunity to develop skills in Eclipse and XML technologies.
A good level in Java programming and HCI design is expected for the applicants. This project is focused on improving the ergonomy of the BPEL tools.

The subject consists in studying the capability to launch the OW2 EJB3 container on top of the Google Android operating system. A prototype and a tutorial will be provided.
EasyBeans (http://www.easybeans.net)
Android (http://code.google.com/android/)

The subject consists in studying the capability to launch the OW2 EJB3 container on top of the Google AppEngine cloud computing service. A prototype and a tutorial will be provided.
EasyBeans (http://www.easybeans.net)
AppEngine (http://code.google.com/appengine/)

The subject consists in developing a console module (flex based named kerneos) for managing the RSS sources. By default, the feature permits to follow the news on the JOnAS and OW2 Web sites.

JASMINe monitoring provides a monitoring infrastructure for SOA platform with a flex based web interface. The subject consists in providing an android application enabling the user to connect to the JASMINe Monitoring server from its mobile or tabllet, to get the latest alarms, to configure the monitoring, to view some graphs etc .... At first the application will ask the user for getting some parameters in a dedicated interface and then a html based interface can be used.
Iphone (http://developer.apple.com/iphone/)
Android (http://code.google.com/android/)
JOnAS (http://jonas.ow2.org)
JASMINe (http://jasmine.ow2.org)

JOnAS provides a management console (named jonasAdmin) flex based. The subject consists in providing a admin console for android phone or tablet with less capabilities regarding the existing one (more basic). At least the console must provide some monitoring features (state, list of Java EE components, datasource, JMS objects, …) and deployment feature as well.

Hyperic is an extensible administration tool enabling to develop plugins for managing various third parties products. The subject consists in developing the plugin to support the JOnAS application server.
Hyperic (http://www.hyperic.com)

Puppet is a configuration tool enabling to automate the installation and configuration of a distributed platform. The subject consists in providing a puppet template for supporting the configuration of JOnAS application server.
Puppet (http://www.puppetlabs.com/)

Intellij IDEA is a well known IDE providing an alternative to Eclipse or Netbeans development tooling. The subject consists in developing a IDEA plugin for supporting the JOnAS application server with the same features than the JOPE plugin for Eclipse.
IntelliJ IDEA (http://plugins.intellij.net/)
JOPE (http://jope.ow2.org/)

Android is a software stack for mobile devices, the Android SDK provides the tools and APIs necessary to develop Java applications that run on Android-powered devices. Joram defines a specific client library allowing J2ME applications to access the JORAM platform in a JMS 1.1 "way". The goal of this project is to define an Android component allowing the use of Joram/JMS API on Android-powered devices.

The main objective of the SAM extension is to open up the messaging infrastructure to the PHP scripting environment by providing an API that allows the most common and simple messaging operations to be accomplished with a minimal investment.
The objective is to offer a JORAM connection factory for SAM allowing a simple and convenient use of JORAM with PHP scripts.

Hyperic is an extensible administration tool enabling to develop plugins for managing various third parties products. The subject consists in developing the plugin to support the Joram messaging server.

Nagios is an extensible administration tool enabling to develop plugins for managing various third parties products. The subject consists in developing the plugin to support the Joram messaging server.

Xymon (also known as BigBrother) is an extensible administration tool enabling to develop plugins for managing various third parties products. The subject consists in developing the plugin to support the Joram messaging server.

Service4All is a cloud computing project focusing on PaaS (Platform as a Service) layer, which aims at providing a cloud platform mainly for service-oriented software developers. As Service4All is designed to provide services for a large number of users, the resources that a user can consume are usually specified by certain form of explicit or inexplicit contracts (e.g. SLA-Service Level Agreement).  On the one hand, to understand the resource consuming details,  it is necessary to meter memory, CPU and other resource usage of each user at different granularity levels, such as in application instances, service containers, VMs and physical machines. On the other hand, the system should guarantee the resource provisioning so as to meet the user quota and avoid over-provisioning as well.  The goal of this project is to add resource metering and guarantee support to Service4All

Web services are widely known as the building blocks of typical service-oriented applications. The performance of such an application system is mainly dependent on that of component web services. Thus the effective load testing of web services is of great importance to understand and improve the performance of a service-oriented system. However, existing Web Service load testing tools ignore the real characteristics of the practical running environment of a web service. Service4All is a cloud computing project focusing on PaaS (Platform as a Service) layer, which aims at providing a cloud platform mainly for service-oriented software developers.  In this topic, the developers are required to build a Testing as a Service Platform for Web Service Load Testing with Service4All to support efficient, realistic, massive concurrent Web Service load testing. With this work, testers can input the configuration information of a test task , and then monitor the test process and download the test report. The test resource provision process is transparent to testers, which can observably simplify the manipulation of testers.

The goal of this project is to detect functional and technological features from software texts as software descriptions, code annotations, user comments, etc., and to locate the relevant code fragments. Such facility can improve effectiveness of software retrieval and ease code-reuse. This project can be implemented as a service module to be plugged into software hosting websites like forge.ow2.org. The development of this project can utilize the mechanisms or services provided by Trustie TSRR. Besides, one of the projects in Trustie TSRR, INFLUX, provide abunbant software profiles crawled from multiple OS forges which can be taken as raw data for the project's analysis task.

SMEP is a platform that provides software measure and evaluation for projects in Trustie and Aliteia. The Key features in SMEP are implemented as Web Service. As BPM is an good solution for platform users customize their measure and evaluation, there will be a mechanism provides workflow design , checking and execution. Orchestra is an BPM solution based on BPEL from OW2 which is a good reference. The technology used to develop this feature could be either ODE or Orchestra and apache AXIS2.