Java environment / tool for a simple distributed computing task

We generate pdf files with data on the monthly financial balance of tens of thousands of customers. At its peak (100,000 files at the end of the year), the process can take up to 5 days to distribute the load between 5 servers. Workload distribution is a manual process (for example, server 1 generates PDF for clients from 1 to 20,000, server 2 from 20.001 to 40.000, etc.).

We use Java, so we would like to use a Java tool or framework similar to BOINC (BOINC is not Java encoded), which would allow us to:

  • a server that automatically distributes and manages data and code for workers
  • working (simple dual-core desktop computers using free (or not so) processor cycles) receiving the code sent by the server and executing it using data as input, providing some feedback on the status and, finally, doing something with output, for example. sending PDF files to network storage.

At the moment, we do not want to use super powerful omnipotent tools, such as the Globus Toolkit.

¿Do you know any java framework / tools suitable for our needs? Open source will be enjoyable, but we are not abandoning commercial solutions.

+3
source share

Source: https://habr.com/ru/post/1716161/


All Articles