ProActive Parallel Suite, http://proactive.inria.fr/, developed by ActiveEon, http://www.activeeon.com/, provides a powerful innovative Open Source Solution for multi-cores to Enterprise Grid & Cloud computing.

ActiveEon, your dedicated technology partner for accelerating and scaling up your enterprise applications, has strategic relationship with NVIDIA in order to ease the deployment of CUDA Tasks on GPUs and Multi-GPUs.

See ProActive Programming 4.2.1 change log.

See ProActive Scheduling & Resourcing 2.0.1 change log.

The next minor release is scheduled before end of June and will gather new impressive features such as a distributed workflow environment.

Meanwhile, get faster up to speed with ProActive 4.2 by attending ProActive trainings! Next sessions’ schedule.

This major French event was again a good opportunity for the ProActive team to promote the different solutions to build internal cloud using ProActive Scheduling and ProActive Resourcing products.

Furthermore, the OW2 consortium organized a meeting about its cloud initiative recently launched. ProActive was presented as well as a key element of the built solution.

See also the Pitch Minute (in French) performed by the ActiveEon’s founder Denis Caromel!

Come and meet us!

ActiveEon will be present at the exhibition day on the INRIA stands. Further, ProActive integration with Scilab will be discussed during the Scilab user day.

More information at http://www.teratec.eu/gb/forum/index.html.

We really look forward to meeting you at the Ter@tec Forum.

Arnaud Contes, Software Architect at ActiveEon, will introduce to you Cloud computing  principles, give an overview of ProActive Parallel Suite and show how to create your internal Cloud.

Registration is free but mandatory.

We look forward to meeting you all on Friday!

Demos and Discussions with the ProActive team

The Solutions Linux trade show was a great success this year. During the 3 days, ActiveEon’s and INRIA’s  peoples continuously demoed the ProActive features attracting a lot of of interest among the numerous visitors. We were located on the OW2 stand among many booths (Petals, SpagoBI, XWiki, …)  presenting Open Source projects and their supporting company.

Solutions Linux attracted more than 8 000 visitors and 210 exhibitors with a large place to server, distributed and cloud computing with a strong focus on Open Source solutions.

The different talk about ProActive during the SOA & Clusters sessions as well as in the OW2 conference aroused interest and questions:

• Parallel Business Services in a SOA Architecture for a eTravel application
• Multi-Cores to Virtualization and Clouds: Accelerating Demanding Applications and Managing Infrastructures
• Cloud and acceleration of Java applications with ProActive Parallel Suite

ActiveEon on the OW2 stand (on the left)

Denis Caromel's talk during the OW2 conference

Read the Press Release (english, french )

First, ProActive will be discussed during the CS2: The supervision of joint infrastructure panel. Do not hesitate to attend this conference.

Further, ActiveEon will also be present in the Seeds In the Cloud area with a stand.

For more information and detailed agenda see www.datacenter-cloud.com.

We really look forward to meeting you at Solutions CloudComputing.

INRIA presented the ProActive Parallel Suite composed of our three products: a Java middleware for distributed and parallel computing, a resource manager, and a job scheduler. Both the Java middleware and the resource manager implements the GCM Deployment ETSI standards (GCM Interoperability Deployment and GCM Interoperability Application Description).

The INRIA demoed the use of ProActive Resourcing and Scheduling on a local cluster and Amazon EC2. When the cluster is overloaded, computing nodes are dynamically gathered from Amazon EC2.
The test suite and test applications defined by the ETSI have also been implemented and successfully executed on Amazon EC2, PBS, LSF and SSH demonstrating the partial implementation of the GCM standard.

Furthermore, we contributed to the following survey presented to the participants:

• User API: Java API, XML GCM DD/AD descriptors, XML Job descriptors
• Data Management: Input and Output data directories can be shared between processes (remote access)
• Hypervisor & resource management: Supports Xen, VMWare, KVM, Virtualbox, qemu, hyper V. Provides an unified API to manage all the supported hypervisors. Support SSH, PBS, LSF, SGE, OAR and more by using infrastructure specific CLI.
• Resource reservation: Fully configurable through selection scripts. Jobs can ask to run scripts on resources to check if they fulfill their needs.
• Feedback on existing standards:
  • Use ETSI GCM DD/AD XML descriptors to perform static deployments
  • Evaluated OGF JSDL for job submission but too many vendor specific options and does not fulfill all the needs of ProActive Programming/Resourcing
  • Evaluated OVF from the DMTF. It fails to address the main issue of moving images between different hypervisor without human intervention.
  • Looking for a common API for managing hypervisors. Currently, proprietary APIs must be used.