Projekte im BibTeX-Format



\germanTitle{Steer-by-Wire und CarRing II}

\englishTitle{Steer-by-Wire and CarRing II}


\member[Leader]{Prof.\ Dr.\ Harald Richter}


\member[Project Staff]{Dipl.-Inf.\ Marcel Wille}


\member[Project Staff]{Dipl.-Inf.\ Christian Asam





\partner{Dr.\ Sergej\ Alexejew}

{Tambow University of Technology, Russia}




In future steer-by-wire systems for cars, the front wheels are steered in real-time by means of two electric motors


according to the rotation angle and torque of the steering wheel. There is no mechanical connection between steering wheel


and front wheels. This is accomplished by a distributed mechatronic system with a field bus as commmunication means. However,


field busses only have layer 1 and 2 in the ISO 7-layer model. We propose instead a dedicated, fully-fletched, real-time


computer network called CarRing II. CarRing II exhibits functions from all 7 ISO layers and aims at general


intra-car-communications by means of a system’s perspective. It optimally supports future x-by-wire driver-assistance systems


with steer-by-wire as a special case. Its 4 main goals in comparison to field busses are better reliability and real-time


capability as well as better usability and effectiveness. The innovations are in the chosen application itself and in the way


we want to achieve the 4 goals for CarRing II. CarRing II is based on optical plastic fibers arranged in rings that transmit


data with 1 Gbits/s. It has a new medium access with fairness, livelock- and deadlock-avoidance, guaranteed packet latency


and high-bandwith efficiency. It allows for automatic car-wide routing, authentification, authorization, common data-exchange


format and a new programming model with distributed registers and remote interrupts. By this features and its


quality-of-service for data transfers, it is unique in intra-car-communication even compared to its closest competitor the


AUTOSAR (AUTomotive Open System ARchitecture) middleware for interoperable communication between automotive electronics.



\duration{since 2005}{}




\funding{TUC, MWK Hannover, Verein von Freunden der TUC, DAAD}{}{}


\reference{M. Wille, H. Richter, C. Asam, Beyond FlexRay - A Survey of CarRing II, Proc. ATAC Design Forum on Automotive


Bus Systems @ International Solid-State Circuits Conference ISSCC 2007, San Francisco, 11-15. Feb. 2007.}

\reference{C. Asam, M. Wille, H. Richter, Carring II: A Real-Time Computer Network as Successor for Flexray?, Proc.


Embedded Systems Conference at the Electronica Exhibition, Munich, 14.-15. Nov. 2006}







\germanTitle{D-Grid-Teilprojekt Medigrid}

\englishTitle{D-Grid Subproject Medigrid}


\member[Leader]{Prof.\ Dr.\ Harald Richter}


\member[Project Staff]{Dipl.-Inf.\ Dietmar Sommerfeld}



\partner[Main Leader]{Prof.\ Dr.\ Oswald Haan}

{GWDG, Göttingen, Germany}


\partner{as well as many other scientists from more than 11 German institutions}{}





Motivation for Grid Computing

Grid computing is the modern key technology to provide the means to solve very large-scale computational problems by using


distributed heterogeneous resources. It is a special kind of distributed computing with a focus on the use of computing and


data resources across existing administrative domains. Grids provide a new quality of service in resource sharing and problem


solving. They allow the utilization of resources at a large number of distributed sites belonging to different organizations.


Within a Grid all members form a new administrative domain called Virtual Organization (VO).

If an application needs more resources than a single computing center can provide the application be gridified, i.e.


distributed to several sites. Gridification is the process of enabling an application for execution on a Grid

Gridfication allows not only to execute an application that is too large for one site. It reduces additionally the execution


time of smaller jobs that would fit into the site by acquiring resources from outside that were not accessible before.


Usually resources in computing centers not 100% utilized. Such unused resources are made available to the grid. Additionally,


computational loads are never constant. With grid computing an overloaded computing center can migrate jobs to sites with


lesser load. The latter two facts increase the computing center's efficiency. Finally, if a machine in a computing center is


down job migration ensures reliability.

The number of resources made available via Grid will increase. Additionally, many public research institutes will not be


funded enough to have large local clusters for their own. Instead, computational projects will be supported that buy


computing power from a grid.


Grid computing by MediGRID

MediGRID is part of the German e-Science initiative D-Grid of the BMBF. The aim of the project is to provide a community Grid


for researchers in the fields of medicine, biomedical informatics, and life sciences. Four types of pilot applications were


selected for the first phase of the MediGRID: bioinformatics, image processing, biomedical ontology, and clinical research


applications. Together with the GWDG, we are engaged in the subproject about bioinformatics.



\duration{since 2005}{until 2008}




\funding{MWK Hannover, Gesellschaft für wissenschaftliche Datenverarbeitung Göttingen (GWDG) by means of BMBF project


DGrid, subproject MediGrid}{}{}







\germanTitle{DEISA-Teilprojekt Metascheduler}

\englishTitle{DEISA Subproject Metascheduler}


\member[Leader]{Dr.\ Thomas Soddemann}


\member[Leader]{Prof.\ Dr.\ Harald Richter}


\member[Project Staff]{Dipl.-Inf.\ Janko Heilgeist}


\partner{as well as many other scientists from more than 13 European institutions}{}





With the advent of grid computing, a special form of distributed computing, computing resources of many distributed


centers are available to their customers which are themselves in some projects distributed all over the world. In order to be


able to efficiently offer resources to customers, the necessity grows to balance resource requests across grid


infrastructures automatically.

Currently, with a few exceptions grid infrastructures only offer limited support for brokering resources. Grid middlewares


like UNICORE and the Globus Tool Kit require the user to make the decision on specifying the site which may be offering




DEISA is the Distributed European Infrastructure for Supercomputing Applications we are participating with as subcontractor


of the Max-Planck Computing Center Garching. DEISA is a research project of the European community (FP6) and formed by eleven


of the top European computing centers. Hence, it is much more than just another Grid project. DEISA offers its customers the


unique opportunity to access supercomputing resources hardly available elsewhere. While most of the computing resources are


used at single sites, there are projects which can make use of resources allocated at multiple sites.

Due to nonexisting middleware, currently the resource allocation for such requests is performed manually. Those deficiencies


in automated cross-site resource allocation lead to a few attempts of improving the situation. E.g., the Gridway project


tries to address the problem especially in a Globus Tool Kit context. The Platform project offers a proprietary solution with


LSF Multicluster. All these attempts have some disadvantages our approach will try to circumvent. The core of our approach is


a distributed meta-scheduling architecture which allows the migration of jobs between the participating resource providers of


a grid-like infrastructure with the aim of improved resource utilization, load balancing, and turn over times. The approach


makes use of Peer-2-Peer-based algorithms and multicriteria decision algorithms.



\duration{since 2006}{until 2008}




\funding{Max-Planck Computing Center Garching (RZG) by means of EU Project DEISA, subproject Metascheduler}{}{}






Kontakt  Suche  Sitemap  Datenschutz  Imprint
© TU Clausthal 2018