Java Parallel Processing Framework (JPPF) version 1.4 available

Home

News: Java Parallel Processing Framework (JPPF) version 1.4 available

  1. In this version:
    • Nodes can now be restarted or shutdown remotely from the administration console or by API.
    • Tasks can be executed locally by the client, with automatic load-balancing between local and remote execution.
    • JPPF nows take full advantage of multi-core multi-CPU hardware
    • the exact same code can be executed purely locally, in a pure distributed way, or in a mix of both
    • the overhead of distributed execution disappears for computationally small tasks
    • The nodes security policy can be downloaded from the server for an even easier deployment.
    • The nodes collect the CPU time used by the tasks
    Once again, we thank the JPPF users community for inspiring these great features and contributing to improve the product. About JPPF: JPPF is an open source Grid Computing platform written in Java that makes it easy to run applications in parallel, and speed up their execution by orders of magnitude. JPPF has many features:
    • a JPPF grid can be up and running in minutes
    • highly scalable, distributed framework for the execution of Java tasks
    • leverages JCA 1.5 to integrate with leading J2EE application servers
    • easy programming model that abstracts the complexity of distributed and parallel processing
    • graphical and programmatic tools for fine-grained monitoring and administration
    • fault-tolerance and self-repair capabilities ensure the highest level of service and reliability
    • a set of fully documented sample applications, demonstrating the use of JPPF on real-life problems
    • very flexible open-source licensing
    • and more ...
    More information is available on the JPPF.org web site.
  2. Congratulations, Great to see a solid solution in Java. Can someone please tell us how does this compare to MPI and PVM for example (in terms of job submission, programming model and performance)? It has always been a pain for me to develop MPI programs. I just hate it. Thanks
  3. Great to see a solid solution in Java.
    Thanks for the feeback, this always sheds sunshine on my day :)
    Can someone please tell us how does this compare to MPI and PVM for example (in terms of job submission, programming model and performance)?
    It has always been a pain for me to develop MPI programs. I just hate it.
    Welcome to the club. The complexity of standards like MPI, their weak integration (or lack thereof) with Java, are exactly why I started JPPF in the first place. I never bothered to perform any comparison between JPPF and MPI, so I can't really address that question. Not sure how objective I could manage to be, anyway... -Laurent
  4. Parallel Java[ Go to top ]

    As I see it, JPPF is really more a generic grid platform in 100% Java, i.e. it does job submission, load balancing, some monitoring&management. If you want to do parallel programming it's more a 'do it yourself' approach. You can do parallel programming with it but how to divide a program into subtasks and reassemble the results from them is up to you. For this I think Parallel Java http://www.cs.rit.edu/~ark/pj.shtml is better: it gives you MPI like functionality but also openMP like (for multithreading on one node) functionality. The grid management functionality of PJ is rather primitive (job submission is all done via ssh, no load balacing) and could be improved for example by taking it's own job management out en implement that via JPPF (so integrate PJ on top of JPPF). This should be feasible since both are open source & Java.
  5. Could you kindly explain what are the differences between JPPF and GridGain? Advantages/disadvantages of JPPF vs Paremus/InfiniFlow and/or GigaSpaces?
  6. Hello Konstantin,
    Could you kindly explain what are the differences between JPPF and GridGain?
    There's a lot of differences, and I do not know them all. I suggest to take a look to this previous thread on TSS, where some of the major ones are listed and discussed by the teams of both products. I think both are great and provide a rare thing nowadays: the ability to choose between solutions. I invite you to try them for yourself and find out which one best fits what you need.
    Advantages/disadvantages of JPPF vs Paremus/InfiniFlow and/or GigaSpaces?
    I don't think it makes any sense to compare JPPF with InfiniFlow or GigaSpaces. JPPF is a computational grid platform that essentially addresses a class of problems called embarrassingly parallel problems (Wikipedia). The other products you mention are more like full-fledged application servers and address a much broader set of features and requirements. -Laurent
  7. Sorry I forgot the TSS link I mentioned: http://www.theserverside.com/news/thread.tss?thread_id=47941
  8. Congratulation!! I know this is really helpful for distributed computing and specially clustering and replication. -Deepal http://blogs.deepal.org
  9. Laurent I noticed that you have plans to integrate with GigaSpaces in one of your future releases. http://jppf.org/presentation.php?current=12 Konstantin Note also the integration of Fura Grid with GigaSpaces: You can see the details of this integration here. So as Laurent mentioned GigaSpaces, JPPF, Fura and other Grid solutions are pretty complementary. In fact its pretty common to have GigaSpaces and a compute-grid solution such as DataSynapse or Platform in the financial industry to for a conducting compute intensive analytics. Nati S. GigaSpaces
  10. Hello Nati,
    I noticed that you have plans to integrate with GigaSpaces in one of your future releases.
    Thank you for the reminder :) I have officially registered it as a JPPF feature request: 2074492 - Integration with GigaSpaces, which means that we will be actively working on it from now on. -Laurent