Embedded systems (ES) are application-specific, dedicated computing devices characterized
by an intensive interaction with their physical environment. They observe different
physical quantities through their sensors, provide complex computation and react to
the world via their actuators. Networked ESs may fusion their information (e.g. sensor
information) and cooperate in the execution of a joint collaborative task. The extreme
communication speed and widespread access to the Internet allows a reformulation of
the role of embedded systems. Our vision is a huge network composed of general purpose
embedded systems, which share their knowledge, computation resources and interaction
capabilities (actuators) to serve as execution platforms for many tasks allocated
at run-time. Emerging user demands and tasks are dynamically decomposed and downloaded
to the embedded systems on-the-fly without interrupting the already scheduled tasks.
This requires high level modeling and knowledge description, appropriate configuration
monitoring and resource management, and virtualization of both the embedded systems
and sensors. Cloud computing is a favorite candidate for the provision of the extra
computing power, storage and knowledge for moving the system design/deployment task
from the traditional pre-operation design phase to the online operational phase.