4. Placement

In this section, we briefly present the state-of-the-art in methods for placing tasks on the various nodes of a distributed system. Classically, the problems of assigning tasks to the different processors in a distributed architecture fall into two categories:

• static placement: this involves highlighting a distribution of tasks for a given hardware architecture. The allocation procedure takes place before the application is started, and the result is fixed for the life of the application. This type of placement requires a thorough knowledge of hardware and software architectures;

• dynamic placement: this type of problem arises during application execution. Specifically, it may be necessary to reassign tasks when the software architecture is modified (i.e. when a task is created or deleted) or when the hardware...