JPPF makes it easy to parallelize computationally intensive tasks and execute them on a Grid.
JPPF 4.2 brings simpler and more powerful client-side APIs, greater scalability options, emphasis on concurrent jobs in the grid and much more.
What's new in JPPF 4.2
Simplification of the client APIs
- Cancelling a job, as well as getting or monitoring its results can now be done from the job itself, in a much simpler way
- Consequently, the TaskResultListener API and its JPPFResultCollector implementation are now deprecated and superseded by the job listeners API
- JPPFJob now implements Future<List<Task<?>>>
Greater focus on connections pools
- Each connection pool now manages an associated dynamic pool of JMX connections, instead of having one JMX connection per JPPF connection. This results in much less threads created on the client and server sides and increases scalability
- The JPPFConnectionPool API has been greatly enriched, for an easier and more powerful handling of its capabilities
- Similarly, the API to explore connections pools is now broader and more flexible
Emphasis on submitting jobs concurrently
- a new section of the documentation is dedicated to parallel job execution
- a new dedicated sample illustrates the patterns explored in the documentation
New execution policies
- The new IsInIPv4Subnew policy filters nodes based on their membership in one or more IPv4 subnets
- The IsInIPv6Subnet policy plays the same role for IPv6 addresses
Control of tasks resubmission
The maximum number of times a task can schedule itself for resubmission is now configurable at the task level, in addition to the job level.
The JPPF tutorial was updated to account for the greatest and latest features.