Chair of automation / computer science

Dynamic Reconfiguration

The dynamic reconfiguration of hardware stands for the change of hardware while the system is operating. It makes an adaption to different conditions possible. For instance, an improved use of communication networks in process control engineering can be achieved: Many networks reveal the characteristic that connections between specific communication partners show a smaller latency than others. These localities (connections with a small latency) can be used for those communication partners who communicate intensively with each other. If several applications run in parallel on the system and a part of applications change, this can implicate that after the change, new partners with intensive communication cannot use a connection of small latency. Then, a dynamic reconfiguration of the network can help.

Open questions are, when does a reconfiguration make sense and which network architecture is to be used at what time. In this context, the varying spatial traffic distribution caused by changing applications plays an important role, and thus, the application area. Another question is, what happens with the messages that are already within the network at the beginning of the reconfiguration. Switching and routing must also be part of the investigation. New Switching and Routing strategies must be developed. A form of reconfiguration management must be added to the system, which is possible in the context of a hardware operating system.

More information about #UniWuppertal: