The systems development team drives innovation in embedded operating system functionality with a particular emphasis on virtualization. This work harnesses powerful embedded processors (32- and 64-bit CPUs containing multiple cores) to provide features such as isolation (allowing rapid updates or wholesale changes to the system) without disrupting the functionality of the past. Using this approach, companies can deploy improvements to embedded systems far more quickly.
Research Areas
Embedded virtualization – With the availability of high-performance multi-core CPUs, embedded systems manufacturers are looking for new ways to exploit the increased compute power. Virtualization allows manufacturers to not only consolidate multiple embedded systems onto a single system-on-chip (SoC), it also allows them to streamline their systems engineering processes and to engage in new business models. By using embedded virtualization solutions that fulfill embedded requirements (such as deterministic real-time behavior, prioritized fast boot of timing-critical partitions, minimal CPU and memory footprint, etc.), embedded systems builders can design new systems that go beyond the capabilities of existing embedded solutions.
Application Domains
Manufacturing – With virtualization, programmable controllers can run on virtualized and shared hardware. This leverages multicore hardware and enables the same functionality on a smaller number of boxes.
Automotive – Different applications can run in parallel while being strongly isolated against each other. Safety critical and normal application can be hosted on the same hardware independently. Update cycles can be decoupled and customer applications can be deployed without impacting critical automotive systems.
Health care – The separation of critical functions to control the medical device (e.g. CT scan) and the visualization application is the most important aspect of virtualization in the healthcare domain. This enables application developer to use real-time environments to control the device and common application platforms (e.g. Microsoft Windows) for the visualization, configuration and use of the medical device.
- Andreas Lachenmann, Ulrich Müller, Robert Sugar, Louis Latour, Matthias Neugebauer, and Alain Gefflaut, Programming Sensor Networks with State-Centric Services, in Proc. of the 6th IEEE International Conference on Distributed Computing in Sensor Systems (DCOSS '10), Springer Verlag, June 2010
- Andreas Lachenmann, Venkatesh Bommasandra Sadasiva, and Frank Siegemund, An Operator Placement Algorithm for Complex In-Network Processing, in Proceedings of the Sixth International Conference on Networked Sensing Systems (INSS 2009), IEEE, 17 June 2009
