ProActive 4.1, Open Source Java Grid Framework, Available


News: ProActive 4.1, Open Source Java Grid Framework, Available

  1. ProActive Parallel Suite developers are pleased to announce the availability of ProActive Parallel Suite® 4.1, the latest release of the high-end professional open source Java middleware for parallel and distributed computing. ProActive Parallel Suite provides: • ProActive Scheduling: A configurable and powerful job scheduler that helps you to benefit from a higher computing power without having to invest in new hardware resources. It runs everywhere, effectively leverages heterogeneous non dedicated hardware and can be set up and used by everyone. • ProActive Programming: A unique and easy to use library for parallel and distributed programming that makes accelerated and powerful applications within everyone’s reach. Thanks to a comprehensive set of efficient base structured patterns everyone can quickly build a parallel application! • ProActive Optimizing: A large set of graphical control and monitoring tools to optimize workload management, infrastructure use and accelerate further your applications It can be freely downloaded at Major improvements in ProActive 4.1 are the dynamic policy-based resources acquisition and an improved resources allocation, a new javascript console, a more flexible deployment framework with additional configuration parameters and full http support. Find more on ProActive new features here: Visit us at JavaOne (booth 628)!

    Threaded Messages (2)

  2. License?[ Go to top ]

    I cannot find a guess on licensing. Some googling brought me to both LGPL and GPL - which is quite different. Also why should be proactive is better (or different) from gridgain and hadoop? Thanks, Raffaele
  3. Re: License?[ Go to top ]

    Hi Raffaele, ProActive is released under GPL v2 (or later) license since version 3.9. Other Licenses are available from INRIA or ActiveEon, for instance for OEM partners. Regarding your question about Hadoop and GridGain, both of them are map-reduce "only" frameworks, with Hadoop being more data-oriented than GridGain. Basically, both of them divide applications into many small blocks of work, and place them on compute nodes managed by the framework. ProActive Scheduling, part of ProActive Parallel Suite, provides such a feature. However, unlike Hadoop and GridGain, the application you submit to the Scheduler is not restricted to map-reduce processing. Firstly, it can be any set of tasks, ordered or not (bag of tasks, DAG, pipeline,...) using either an XML-based definition or the straight-forward Java API. Secondly, the submitted application can be also any ProActive Programming application, for instance a communicating SPMD application. One other difference is the resources management framework : ProActive Scheduling comes with a powerful graphical tool for resources control and monitoring. It allows interactive as well as time-based or workload-based resources allocation. Moreover, the deployment of computing nodes can be done through connection protocols like RSH or SSH, job submissions to schedulers like PBS, LSF, SGE, Windows HPCS, IBM LoadLeveler or Clouds (Amazon EC2), or accessing computing nodes embedded within J2EE servers. Cheers, Arnaud