How does the application placement controller work with Compute Grid in particular, dispatching jobs and endpoint selection?
When the application placement controller and Compute Grid are configured to work together, Compute Grid delegates the task of endpoint selection to the placement controller. When jobs are submitted, Compute Grid asks the placement controller to select an endpoint on which to run the job. The information that Compute Grid includes in this request to the application placement controller includes the class of the job, the completion time goal, and the nodes, clusters, or servers where the job is permitted to run. When the application placement controller has selected an endpoint, that information is passed back to Compute Grid , which is responsible for starting the job. When the application placement controller selects the endpoint, it considers the completion time goal for the job, available resources in the system, the execution profile of previous jobs in the same class, and other work, including both transactional and batch work, that the system needs to handle. The application plac