Current Projects
ToleranceZone
A Fault Tolerant Middleware Idioms based on Self-Stabilizing Techniques
This research project pursues the goal of furnishing essential operations of wireless sensor networks with fault tolerance based on self-stabilizing algorithms and thus to lay foundation for a permanent and uninterrupted operation of these networks. The project will be based on two fundamental assumptions. Firstly, fault tolerance is not an add-on to an existing middleware, but a recurrent theme throughout the design of the software infrastructure that glues together all components. Secondly, fault tolerance must be a self-organizing property.
The latter assumption demands for a highly decentralized mode of operation. The project will substantiate the claim that the employment of self-stabilizing algorithms will bring about the same degree of fault tolerance, as this can be achieved with state of the art middleware platforms. We will evidence that with the new approach the consumption of resources is considerably reduced and that adaption to new types of errors is an inherent feature. The quantitative analysis of these claims will be carried out through a comparison of existing middleware platforms and a prototypical implementation of our approach.
The project is funded by the DFG and runs from October 2011 until September 2013.
4D-Cast
Survilance of highly dangerous technical devices using non invasive self organizing techniques from wireless sensor networks. Example: A gas station
This project aims at the development of an autonomous wireless sensor node which can be applied directly to devices that should be surveyed, due to its small size. Environmental parameters like temperature, movement or hazardous gas leaks can be measured and analysed right there. Only when a disturbance is detected, the gathered data is transmitted to a central unit. The biggest challenge of this project is the development of a robust sensor node that can be used under extreme conditions, including but not limited to danger of explosions, radio disturbances and concussions. Also, the nodes need to be able to survive for a long time, as it will not be possible to exchange nodes and/or batteries frequently.
TANDEM
This project brings together several research institutes and companies under the guidance of the IHP. The aim is to develop an extreme economical scalable TANDEM-processor based radio system for sensorial, actuarial and indicative applications. Our contribution consists of the REFLEX operating system, solutions for powermanagment and over the air reprogramming techniques.
NOVOS
An Operating System for Architectures with Non-Volatile Memory.
This project started in 2012 and is funded by Intel.
Ongoing Projects
REFLEX
REFLEX is a generic event driven OS for embedded devices. Event handlers and control functions are all represented by passive objects that are scheduled preemptively according to an earliest deadline first (EDF) strategy. All sensors, control functions and actuators of a typical embedded control system are represented by objects that can communicate with each other by means of events.
TACO
TACO (Topologies and Collections) is a highly efficient distributed object platform for cluster architectures which is entirely based on C++ templates. TACO provides both template based remote method invocation mechanisms as well as powerful object groups as the basic means for high-level data-parallel programming. Entire distributed objects groups can be created, destroyed, selectively cloned and manipulated by various parallel collective operations. Thus programmers can treat entire object groups similarly to single objects. Consequently, complex replication and consistency problems can be solved with reasonable programming effort.
Finished Projects
COCOS
The Cocos project(2004 - 2008) brought high-level data parallel communication and coordination paradigms found in the world of massively parallel super computers into the world of tiny embedded systems. It was funded by the DFG (Deutsche Forschungs Gemeinschaft, the german National Research Foundation).
BBGrid
Together with research groups of TU Berlin, University of Potsdam and the Hasso-Plattner-Institut (HPI) we operate the Berlin Brandenburg Grid environment. This Grid infrastructure serves as base for our research in the field of Grid computing, resource management and scientific computing.
