Instead of processing a program on one central station, the processing can be distributed on a network of multiple devices. Examples are smart devices that we recently see in our daily lives as in smart homes or even in factories.
In this project, we focus on programs that process sensed data. Accordingly, there are multiple advantages of distributed processing, such as: privacy, lower cost and most important, redundancy. With distributed processing, we can avoid a single point of failure, and this is the main target of this project.
The students will develop a failover mechanism to handle changes in the network. In the scope of this work, the students will be provided with:
- a network of Raspberry Pis for processing
- a framework to distribute the processing.
At the end of the project, the students shall demonstrate the failover mechanism by introducing changes to the network such as shutting down a node that is already running a process (i.e., node failures) or losing the connection between two nodes.