9

I am developing a web-based application to provide some services. This is a scientific application which performs some processing on data uploaded by users. The methods corresponds to this processing actions are written in Java and they are completely independent. Obviously, the data of the different users are also independent.

I'm looking for a technique to scale this computation on a cluster like MPI in C. I'm searching for the same tool in Java.

Thanks.

cartoonist
  • 5,991
  • 6
  • 37
  • 40
  • 1
    I guess setup a HADOOP cluster requires substantial effort, isn't it? – cartoonist Jul 31 '11 at 09:08
  • 1
    There are companies (such as Cloudera) that package up hadoop software for (relatively) easy deployment on EC2. But yes, there is a bit of a learning curve for both using hadoop and managing your clusters. I do not know how it compares to GridGain, as I've never used the latter. – Ray Toal Jul 31 '11 at 18:18

1 Answers1

9

You've got a lot of solutions to scale computings with Java, but the one I prefer, and which has the most straighforward code to begin is Gridgain

You can have a try with Hazelcast which is a data-grid, techniquely, but can be used for remote execution and parallelisation of action on a cluster

If you're searching explicitely for MPI implementation, you should have a look to MPJ

Grooveek
  • 10,046
  • 1
  • 27
  • 37