The major components of the business grid include the command-line interface, Enterprise JavaBeans (EJB) interface, Web services interface, long-running scheduler, and execution environment.
With the command-line interface, you can submit and control the long-running jobs in the system. The enterprise bean and Web service interfaces provide similar functionality to both Java 2 Platform Enterprise Edition (J2EE) and non-J2EE programs through programmatic interfaces. The administrative console provides a Graphical User Interface (GUI) with which you can configure the long-running scheduler, to define the service policies with which WebSphere Extended Deployment decides where and when to start long-running jobs jobs.
The administrative console allows business grid administrators and submitters to view, manage and perform job related actions that include submitting a job, viewing of jobs, canceling or suspending a job, and resuming a suspended job. Through the job management console, you can submit job schedules with a preferred processing time. Also, you can configure job schedules so that they can recur on a specific time of day or week, and so on.
The long-running scheduler accepts, persists, and schedules the execution of long-running jobs. It manages the job database, assigns job IDs, and selects where and when jobs run. As part of performing this function, the long-running scheduler starts and stops instances of dynamic clusters as dictated by the jobs to be run and defined service policies. This is analogous to the function provided by the application placement controller for transactional J2EE applications. The long-running scheduler is a J2EE application that you can make highly available by using application placement controller-provided functionality.
The balancer component is an arbiter of long-running and transactional work within a node group. Due to the nature of long-running work, co-locating it on the same node with transactional work usually adversely impacts the transactional work. The balancer decides which nodes perform transactional work and which are used for long-running work. These decisions are based on a number of factors, including how well the service policies for the two types of work are met.