The scheduling and execution of bag-of-tasks applications (BoTs) in Clouds is performed on sets of virtualized Cloud resources that start being exhausted right after their allocation disregarding whether tasks are being executed.
In the case of the FPGA, the percentage of time is obtained for each module separately, when the whole system is working, all the tasks are being executed in a pipeline.
To sum up, for the optimal production deployment, one must consider the amount and the time profile of tasks which are being executed with respect to the thread per core ratio which must be adjusted accordingly; basically, high ratio for many short tasks and inversely for long running tasks.
If there are no more tasks, it means that all job tasks are being (or have been already) executed.
The interrupt controller provides a mechanism for a hardware resource in PSoC mixed-signal array devices to change program execution to a new address without regard to the current task being performed by the code being executed.
In such a case, the scheduler and the data manager must select (i) the data replicas to be used, (ii) the site where these data will be gathered for the task to be executed, and (iii) the routing paths to be followed; this is assuming that the selected datasets are transferred concurrently to the execution site.
The job of the task to be executed is selected in lines 2 3 by function chooseJob (Algorithm 2).
From that, how fast a task can be executed can be derived by using the clock frequency.
The second (liveness property) says that every task will be executed if there is at least one correct resource capable of executing it, i.e., no task will be waiting forever to be executed.
Based on the different tenant- and provider-specific policies, the middleware decides where in the multi-cloud a task will be executed or data will be stored.
In the last considered case, on Figure 9, dynamicity is not in which task will be executed next but in how many instances of a same task will be executed.
