We are pleased to announce that our JobScheduler job network solution has successfully completed its first tests in a customer's environment.
This development is more suited for more complex tree-like networks than our eventing and sync point solutions: it incorporates all objects in the network and not just critical sections such as decision and convergence nodes. A typical JobScheduler job network is shown in the diagram:
Larger versions of the graphic: PDF PNG
Each node in JobScheduler job networks is represented by an order with pre and post processing. Progression through the network is controlled by parameters such as successor, predecessor and script to be executed.
The use of both successor and predecessor parameters provides greater flexibility in defining the job network than solutions that only allow one method of definition.
Job networks are currently only able to execute shell scripts. These are stored in their own files and separately to the job network objects. This means that scripts and the job network objets can be seperately integrated in a versioning system and that scripts can be reused in different networks or at different points in a network.
The JobScheduler job network solution is Java based, meaning that it will run on all operating systems that a Java virtual machine can run on.
This development was partly financed by the customer and currently (Nov. 2013) has a pre-release status. We will keep you informed with updates in this important JobScheduler development.
