Institut für Parallele und Verteilte Systeme - Universität Stuttgart
Institut für Parallele und Verteilte Systeme - Universität Stuttgart
Institut für Parallele und Verteilte Systeme - Universität Stuttgart
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
<strong>Institut</strong> <strong>für</strong> <strong>Parallele</strong> <strong>und</strong><br />
<strong>Verteilte</strong> <strong>Systeme</strong><br />
Annual Report<br />
2009/2010<br />
<strong>Institut</strong>e of Parallel and<br />
Distributed Systems<br />
www.ipvs.uni-stuttgart.de
Preface<br />
Celebrating the 20th anniversary of our institute was defi nitely the most pleasant<br />
event in the period covered by this report. On June 19, 2009 we organized a colloquium<br />
celebrating 20 years of exciting research at IPVS. We had the pleasure to introduce<br />
three outstanding researchers to give talks at this colloquium: Dr. J. Cheng (IBM<br />
Fellow <strong>und</strong> Vice President, IBM Almaden Research Center, USA), Prof. T. Fukuda<br />
(Dep. of Micro-Nano Systems Engineering, Nagoya University, Japan) and Prof. U.<br />
Ramachandran (College of Computing, Georgia Tech, USA). Many thanks for their<br />
most inspiring talks! We also would like to thank for the kind addresses given at the<br />
colloquium, in particular for those of Prof. A. Reuter, the fo<strong>und</strong>er of the institute, and<br />
Dr. H. Knorr, who supported the development of IPVS over the years.<br />
While in 2008 our robot soccer team took a remarkable fourth place in the World<br />
Championship, it was the winning team of both the Robocup World Championship<br />
2009 and Robocup German Open in 2009. We congratulate the Roboter Fußball<br />
Club (1. RFC) <strong>Stuttgart</strong> and his coach Prof. Levi on this excellent performance.<br />
During the past year members of IPVS (co-)organized a number of scientifi c<br />
events, including International Workshop on Quality of Context (QuaCon), ITPC<br />
(IBM Technology Partnership Center) Colloquium on Cloud Computing and Nexus<br />
Industry Colloquium, which all attracted more than 100 participants.<br />
Board of Directors:<br />
(from left )<br />
Dipl.-Inf. Michael<br />
Matthiesen (IFS),<br />
Prof. Dr. Bernhard<br />
Mitschang (AS),<br />
Prof. Dr. Sven Simon<br />
(PaS),<br />
Prof. Dr. Kurt<br />
Rothermel (VS; Head<br />
of <strong>Institut</strong>e),<br />
Prof. Dr. Paul Levi<br />
(BV),<br />
Prof. Dr. Marc<br />
Alexander Schweitzer<br />
(SgS)<br />
1
2<br />
In October 2009, Prof. Mitschang was awarded the IBM Faculty Award for his<br />
work on “Acceleration Technologies for Interactive Advanced Analytics to Support<br />
New Intelligence for a Smarter Planet”. Th e IBM Faculty Award is a highly recognized<br />
international award for excellent achievements in science and teaching as well as for<br />
building a resilient cooperation between research and industry.<br />
Two new colleagues joined IPVS, both strengthening the fi eld of simulation. On<br />
May 15, 2010 Prof. M. A. Schweitzer became head of the chair “Simulation of Large<br />
Systems”. His research is in the fi eld of Scientifi c Computing. Prof. M. Bader took<br />
over the Juniorprofessor position “Simulation Soft ware Engineering” on December 1,<br />
2009. He is conducting research in the area of effi cient algorithms and soft ware development<br />
in scientifi c computing. We warmly welcome both of them and wish them the<br />
very best for their work at IPVS.<br />
Finally, we would like to thank our partners and friends in industries and academia<br />
for their continuous support. We also appreciated the great eff ort of our staff to<br />
bring IPVS forward!<br />
For the Board of Directors: Prof. Dr. rer. nat. Dr. h.c. Kurt Rothermel<br />
Head of the <strong>Institut</strong>e IPVS
Annual Report 2009/2010<br />
(April 1, 2009 - March 31, 2010)<br />
<strong>Universität</strong> <strong>Stuttgart</strong><br />
<strong>Institut</strong> <strong>für</strong> <strong>Parallele</strong> <strong>und</strong> <strong>Verteilte</strong> <strong>Systeme</strong> (IPVS)<br />
<strong>Universität</strong>sstraße 38<br />
70569 <strong>Stuttgart</strong><br />
Deutschland<br />
www.ipvs.uni-stuttgart.de<br />
Published by: Board of Directors of the IPVS<br />
Edited by: Ute Gräter<br />
Michael Matthiesen<br />
Stefanie Palmer<br />
Design by: Michael Matthiesen<br />
Every reproduction - even in extracts - only with the<br />
written consent of the publisher.<br />
© IPVS, 2010
Content<br />
Preface 1<br />
Content 5<br />
1 Organisational Structure 7<br />
2 Research Activities 9<br />
2.1 Applications of Parallel and Distributed Systems 9<br />
2.2 Applied Computer Science - Image Understanding 17<br />
2.3 Parallel Systems 23<br />
2.4 Simulation of Large Systems 30<br />
2.5 Distributed Systems 31<br />
2.6 Joint Activities 40<br />
2.7 Infrastructure 42<br />
3 Teaching 45<br />
4 Dissertations and<br />
Postdoctoral Qualifi cations 47<br />
Dissertations<br />
5 Publications 55<br />
Books / Book Chapters<br />
Articles in Journals<br />
Articles in Proceedings<br />
Others<br />
6 Events 65<br />
7 Professional Activities<br />
Memberships<br />
Program Committees<br />
69<br />
5
Organisational<br />
Structure<br />
Management<br />
Head of <strong>Institut</strong>e<br />
▶ Prof. Dr. rer. nat. Dr. h.c. Kurt RRothermel<br />
Deputy Head of <strong>Institut</strong>e<br />
▶ Prof. Dr.-Ing. habil. Bernhard rd Mitschang<br />
Mitsch<br />
Board of Directors<br />
▶ Prof. Dr. rer. nat. habil. Paul Levi evi<br />
▶ Dipl.-Inf. Michael Matthiesen<br />
▶ Prof. Dr.-Ing. habil. Bernhard Mitschang its<br />
▶ Prof. Dr. rer. nat. Dr. h.c. Kurt Rother Rothermel<br />
▶ Prof. Dr. rer. nat. Marc Alexander xander Schwe Schweitzer<br />
▶ Prof. Dr.-Ing. Sven Simon<br />
6
1 Organisational Structure<br />
Departments<br />
Applications of Parallel and Distributed Systems<br />
Anwendersoftware (AS)<br />
▶ Prof. Dr.-Ing. habil. Bernhard Mitschang<br />
email: Bernhard.Mitschang@ipvs.uni-stuttgart.de<br />
Tel: +49-711-7816449 - Fax: +49-711-7816424<br />
Applied Computer Science - Image Understanding<br />
Praktische Informatik - Bildverstehen (BV)<br />
▶ Prof. Dr. rer. nat. habil. Paul Levi<br />
email: Paul.Levi@ipvs.uni-stuttgart.de<br />
Tel: +49-711-7816387 - Fax: +49-711-7816250<br />
Parallel Systems<br />
<strong>Parallele</strong> <strong>Systeme</strong> (PaS)<br />
▶ Prof. Dr.-Ing. Sven Simon<br />
email: Sven.Simon@ipvs.uni-stuttgart.de<br />
Tel: +49-711-7816450 - Fax: +49-711-7816250<br />
Simulation of Large Systems<br />
Simulation großer <strong>Systeme</strong> (SgS)<br />
▶ Prof. Dr. rer. nat. Marc Alexander Schweitzer<br />
email: Marc.Alexander.Schweitzer@ipvs.uni-stuttgart.de<br />
Tel: +49-711-7816465 - Fax: +49-711-7816248<br />
Distributed Systems<br />
<strong>Verteilte</strong> <strong>Systeme</strong> (VS)<br />
▶ Prof. Dr. rer. nat. Dr. h.c. Kurt Rothermel<br />
email: Kurt.Rothermel@ipvs.uni-stuttgart.de<br />
Tel: +49-711-7816434 - Fax: +49-711-7816424<br />
Infrastructure<br />
Infrastruktur (IFS)<br />
▶ Dipl.Inf. Michael Matthiesen<br />
email: Michael.Matthiesen@ipvs.uni-stuttgart.de<br />
Tel: +49-711-7816414 - Fax: +49-711-7816248<br />
7
Research<br />
Activities<br />
8
2 Research Activities<br />
2.1 Applications of Parallel and<br />
Distributed Systems<br />
<strong>Universität</strong> <strong>Stuttgart</strong> - IPVS<br />
Anwendersoftware<br />
<strong>Universität</strong>sstraße 38<br />
70569 <strong>Stuttgart</strong><br />
Germany<br />
Research Focus<br />
Prof. Dr.-Ing. habil. Bernhard Mitschang<br />
� Bernhard.Mitschang@ipvs.uni-stuttgart.de<br />
� +49-711-7816449<br />
� +49-711-7816424<br />
Th e importance of information in business and also in our daily life has drastically<br />
increased in the last few years. We are constantly challenged by escalating variety, velocity,<br />
and volume of data. Our <strong>und</strong>erstanding of information management is about<br />
creating and enabling business by integrating, analyzing, optimizing, and storing heterogeneous<br />
types and sources of information throughout its life-cycle and in an effective<br />
manner. Th is is refl ected in the diff erent research projects that are conducted<br />
in our department of Applications of Parallel and Distributed Systems. Th e research<br />
work of the department can be categorized into the following three working areas:<br />
▶ Data and Metadata<br />
F<strong>und</strong>amental prerequisites for any eff ective information management comprise<br />
effi cient data and metadata management by means of enhanced database technology.<br />
Explicit metadata management further enables a semantic treatment, interoperability<br />
as well as a code generating approach as applied by model-driven<br />
engineering.<br />
▶ Content and Semantics<br />
Content management defi nes the complementing facet of information management.<br />
Content is <strong>und</strong>erstood to be unstructured or semi-structured and thus demands<br />
for diff erent storage, search, and management techniques as compared to<br />
9
10<br />
the well-structured data management area. Approaches that bridge this gap try to<br />
address both aspects in order to get the best of both sides.<br />
▶ Information Systems, Applications, and Services<br />
Th e above mentioned technologies enable the development of new techniques<br />
for information systems and associated applications, like process or product data<br />
management, message monitoring and retention systems, or context-based information<br />
systems. Th e service-based approach is seen as a next technology step towards<br />
enhancing fl exibility.<br />
Th e research activities of the department are best characterized by its current involvement<br />
in research projects as described further in the project description section below.<br />
Th e department is involved in two programs as part of the excellence initiative granted<br />
to the <strong>Universität</strong> <strong>Stuttgart</strong> by the German Research Fo<strong>und</strong>ation (DFG):<br />
▶ Th e department coordinates all activities within the topical cluster “Information<br />
and Communication Technologies for Manufacturing” that is part of the Graduate<br />
School GSaME (Graduate School advanced Manufacturing Engineering).<br />
▶ Th e department participates in the research area “Integrated Data Management<br />
and Interactive Visualisation” as part of the Cluster of Excellence “Simulation<br />
Technology” (SimTech).<br />
In 2009 Prof. Mitschang was awarded the prestigious IBM Faculty Award for his work<br />
on “Acceleration Technologies for Interactive Advanced Analytics to Support New Intelligence<br />
for a Smarter Planet”.
Staff<br />
Staff of the Department:<br />
(from left ) V. Rewucki, M. Großmann, C. Lübbe, O. Schiller, A. Brodt, A. Roesler, H. Schwarz,<br />
S. Radeschütz, P. Janowski, J. Minguez, T. Scheibler, Ch. Stach, P. Reimann, S. Silcher, B. Mitschang,<br />
F. Niedermann, N. Cipriani<br />
(staff not shown on the photo) A. Haufl er, M. Haufl er, N. Hönle, M. Jakob, A. Moosbrugger, M. Vrhovnik,<br />
F. Wagner<br />
Research Projects<br />
www.ipvs.uni-stuttgart.de/abteilungen/as/forschung/projekte<br />
DaMT - Data and Multi-Tenancy<br />
Personnel: Oliver Schiller<br />
F<strong>und</strong>ed by: <strong>Universität</strong> <strong>Stuttgart</strong><br />
Duration: since 1.1.2009<br />
Multi-Tenancy is about providing a Soft ware as a Service off ering for multiple tenants<br />
in a cost-effi cient way. Th ereby, an accomplished manageability of the service and a<br />
clear-cut and secure isolation between tenants constitute the key challenges. In this<br />
project, we focus on extending traditional data management concepts and systems by<br />
Multi-Tenancy.<br />
11
12<br />
BIA - Business Impact Analysis<br />
Personnel: S. Radeschütz<br />
F<strong>und</strong>ed By: IPVS, partially Deutsche Forschungsgemeinschaft (DFG)<br />
Duration: since 1.6.2006<br />
For optimizing business processes eff ectively, a profo<strong>und</strong> analysis of all relevant business<br />
data in a company is necessary. Th is project aims to develop an approach to automatically<br />
integrate a company’s process execution data and operational business data<br />
in order to enable a more informative and comprehensive analysis. Th e BIA project<br />
addresses important issues like the development of matching algorithms, of extended<br />
analysis techniques in OLAP and data mining and of process optimization procedures.<br />
GSaME – Cluster E - Data Management and Information Provisioning<br />
in the Virtual Factory<br />
Personnel: S. Silcher, J. Mínguez, M. Jakob, T. Scheibler<br />
F<strong>und</strong>ed By: Graduate School advanced Manufacturing Engineering (GSaME)<br />
Duration: 1.4.2008 - 30.9.2013<br />
Th e real-time factory off ers an intelligent, real-time operational management of factory<br />
processes and resources. Th e shift to real-time factory requires signifi cant advances<br />
in the factory’s information technology infrastructure. Th e main challenge is to design<br />
a comprehensive IT architecture for advanced manufacturing engineering based on a<br />
fl exible solution that integrates all manufacturing environments across the entire product<br />
lifecycle.<br />
SQL4WL - Optimization of Data Processing in Business Processes<br />
Personnel: M. Vrhovnik, H. Schwarz<br />
F<strong>und</strong>ed by: <strong>Universität</strong> <strong>Stuttgart</strong>, IBM (partially)<br />
Duration: 1.1.2006 - 30.4.2010<br />
Enterprises organize their business activities in terms of workfl ows. Data management<br />
is usually an integral part of these workfl ows. In this project, we develop concepts for<br />
the optimization of data processing and data management in such data-intensive workfl<br />
ows. In particular, we cover the appropriate architecture, optimization rules, control<br />
strategies and internal workfl ow representations for an optimization framework.<br />
Query-generating Systems<br />
Personnel: H. Schwarz<br />
F<strong>und</strong>ed by: <strong>Universität</strong> <strong>Stuttgart</strong><br />
Duration: 1.1.2007 - 28.2.2010<br />
Many information systems generate queries in order to access data in heterogeneous<br />
data sources. Starting with a detailed study of the broad spectrum of application areas,
we investigate implementation and optimization concepts for such query-generating<br />
systems. Th is covers the main strategies to generate queries as well as new optimization<br />
concepts, e.g., for heterogeneous data management tasks that are connected by complex<br />
control fl ow.<br />
SFB 627 – Project B1 – Homogenized and Virtualized Model Management<br />
Personnel: A. Brodt, N. Cipriani, M. Großmann, C. Lübbe, C. Stach<br />
F<strong>und</strong>ed by: Deutsche Forschungsgemeinschaft (DFG)<br />
Duration: 1.1.2003 – 31.12.2010<br />
During the second f<strong>und</strong>ing period, this project focuses on advanced data management<br />
and processing infrastructure and concepts for the Nexus world model. Th is includes<br />
NexusDS, a scalable and fl exible data stream processing framework, which off ers support<br />
for application-specifi c operators and data, as well as approaches for integrating<br />
query-response, stream- and event-based interaction patterns.<br />
SFB 627 – Project B5 – Dynamic, Time-Referenced Model Data<br />
Personnel: N. Hönle<br />
F<strong>und</strong>ed by: Deutsche Forschungsgemeinschaft (DFG)<br />
Duration: 1.1.2007 – 31.12.2010<br />
To support data analysis and prognosis, Nexus requires concepts for managing timereferenced<br />
data. Th is includes history servers for storing such data, domain-specifi c<br />
pre-processing algorithms for reducing the data volume of trajectories and support for<br />
interpolation functions.<br />
Content Management as a Service<br />
Personnel: F. Wagner, A. Moosbrugger<br />
F<strong>und</strong>ed by: IBM (partially)<br />
Duration: 1.6.2007 - 31.5.2009<br />
Due to the vast amount of valuable semi-structured and unstructured information<br />
produced, processed and stored today, appropriate Enterprise Content Management<br />
(ECM) is a crucial task for companies. In this joint project with IBM and the University<br />
of Hamburg a service-based infrastructure for ECM is developed which can<br />
dynamically adapt to changing workload situations. Another part of the project was<br />
to integrate the prototype into the IBM eDiscovery Manager as a proof of concept.<br />
13
14<br />
SRC: SimTech - Data Provisioning for Scientifi c Workfl ows<br />
Personnel: P. Reimann<br />
F<strong>und</strong>ed by: Deutsche Forschungsgemeinschaft (DFG)<br />
Duration: 1.5.2008 – 30.4.2011<br />
As part of the Cluster of Excellence “Simulation Technology” (SimTech) this project<br />
primarily aims at off ering a consolidated data management and provisioning abstraction<br />
for scientifi c workfl ows, and in particular for simulations. We develop and compare<br />
appropriate data (de-)provisioning techniques and defi ne strategies for optimization.<br />
SRC: SimTech – A Framework to Capture Digital and Physical Test<br />
Cases<br />
Personnel: P. Janowski<br />
F<strong>und</strong>ed by: Daimler AG<br />
Duration: 1.10.2009 – 30.9.2012<br />
In the automotive industry more and more testing processes are accomplished using<br />
simulation technologies. In order to effi ciently assure the claimed functionality of<br />
components, systems and overall vehicles, an appropriate checking process has to be<br />
defi ned, planned and realized. In this joint project with Daimler AG a framework is<br />
developed to capture both digital and physical test cases. By defi ning unifi ed methods<br />
for data access, computation and storage, this allows to bridge between the digital and<br />
the physical world and vice versa.<br />
Selected Research Projects<br />
SFB 627 – Project B5 – Dynamic, Time-Referenced Model Data<br />
In this project, we focus on research in storage and analysis of time-series data in History<br />
Servers and History Warehouse.<br />
History Servers should be able to store time-series data of diff erent data types<br />
persistently. Intelligent preprocessing of stream data reduces data volume and update<br />
rate, so that it is possible to store histories of stream data in History Servers based on<br />
standard database management systems. However, the preprocessed data gives value<br />
patterns using time-dependent interpolation functions, and the currently used basic<br />
data types as well as the Nexus meta data model are only able to defi ne stationary timedependent<br />
data values. Th erefore we extended our basic data type defi nition, so that<br />
it is possible to defi ne time-dependent interpolation functions for data types like fl oat<br />
and geographic types, where continuously defi ned value patterns make sense, e.g., measurement<br />
values of temperature sensors or positions of mobile objects. We also imple-
mented a History Server prototype based on a DB2 database management system including<br />
the DB2 Spatial Extender for geographic data types. Th e query processing of<br />
this History Server uses bo<strong>und</strong>ing box information stored together with the interpolation<br />
functions for quick data access, and automatically calculates data values using the<br />
interpolation functions.<br />
For the analysis of time series data in the Smart Factory project (SFB 627 Project<br />
D1), we developed a so called Smart Factory Warehouse. To extract, transform,<br />
and load the source data into the warehouse, NexusDS (SFB 627 Project B1) is used.<br />
While interpreting the data as stream data, the extraction process is synchronous, i.e.,<br />
new source data will be transferred to the warehouse immediately.<br />
In addition, the Nexus platform hides the schema heterogeneity of the data sources<br />
in a transparency layer, because temporal and spatial data like trajectories are provided<br />
in the Augmented World Model, a specialized data model for context-aware applications.<br />
Th e Nexus Preprocessing<br />
operator (NPC)<br />
for compressing temporal<br />
stream data as position<br />
histories is useful for the<br />
loading of the warehouse<br />
as well. It transforms single<br />
measurement values of<br />
sensors into interpolated<br />
trajectories and therefore<br />
guarantees, that for every<br />
Wrapper<br />
Smart Factory<br />
Main Computer<br />
Wrapper<br />
Wrapper<br />
Sensors<br />
NexusDS<br />
Data<br />
Warehouse<br />
point in time for which data has to be loaded, there will be a sensor value that can be<br />
loaded into the warehouse. Modeling the uncertainty of compressed trajectory data is<br />
crucial for correct interpretations and analysis.<br />
Content Management as a Service<br />
In today’s businesses it’s getting more and more important to have a highly productive<br />
Enterprise Content Management (ECM) system to cope with an always increasing<br />
amount of structured and unstructured data. Th is data itself must not only be<br />
managed, archived, and retrieved, it’s also desired to make the information available<br />
through search for easier selection of important documents. Th is is also relevant for<br />
compliance scenarios, where in case of a law suit all necessary documents have to be<br />
given to court in a certain time frame.<br />
NPC<br />
Data<br />
fl ow of<br />
the ETL<br />
process<br />
15
16<br />
Another important factor is, that companies nowadays have an increased interest<br />
in effi cient and scalable ECM-services. Th ese services would allow companies using<br />
ECM applications in a utility based, cost-eff ective manner.<br />
In this joint project with IBM Research & Development Germany in Böblingen<br />
and the University of Hamburg a service-based infrastructure for ECM was developed<br />
which can dynamically adapt to changing workload situations. In this joint research<br />
project as sample ECM application an Email Archiving and Discovery solution, like i.<br />
e. IBMs Content Collector, was selected. A prototype was implemented and used to<br />
validate the research results. In the prototype we developed a distributed data model<br />
for this scenario and used known technologies from P2P-systems to create a scalable<br />
system. Furthermore, we integrated an Apache Lucene based search cluster, that manages<br />
itself and that is capable to cope with varying load. Another part of the project<br />
was to integrate the prototype into the IBM eDiscovery Manager as a further proof of<br />
concept.
2.2 Applied Computer Science -<br />
Image Understanding<br />
<strong>Universität</strong> <strong>Stuttgart</strong> - IPVS<br />
Bildverstehen<br />
<strong>Universität</strong>sstraße 38<br />
70569 <strong>Stuttgart</strong><br />
Germany<br />
Research Focus<br />
Prof. Dr. rer. nat. habil. Paul Levi<br />
� Paul.Levi@ipvs.uni-stuttgart.de<br />
� +49-711-7816387<br />
� +49-711-7816250<br />
Image <strong>und</strong>erstanding denotes the multi-sensorial and knowledge based interpretation<br />
of situations, estimations of intentions and predictions of activities of the perceived<br />
objects. Th is division defi nes and develops distributed intelligent systems on macro-<br />
and microscopic scale:<br />
▶ Image Understanding<br />
Th is scientifi c focus is devoted to the context based recognition of situations and<br />
the prediction of the behaviour of the detected objects. Such a cognitive interpretation<br />
of situations is usually distributed and has to be proved on consistency<br />
before it is transferred in to a world model.<br />
▶ Multi-agent systems<br />
Autonomous systems operate <strong>und</strong>er self-control of their internal states, goals and<br />
actions and act usually without human interactions. Such systems are generally<br />
modelled by multiple agents. Representative applications are the development of<br />
autonomous logistic systems, autonomous mobile systems especially autonomous<br />
cars and traffi c management by these cars themselves.<br />
▶ Cognitive and evolutionary Robotics<br />
Cognitive robotic systems are provided with an elaborate cognitive architecture<br />
that cooperates closely with a controlling architecture in order to be able to act<br />
fully autonomously. A top-down approach is used to develop autonomous multi<br />
robot systems like for example in the RoboCup scenario. On a microscopic level<br />
evolutionary robots are developed to investigate collective or swarm intelligence.<br />
▶ Nonlinear dynamics and chaos theory<br />
Th e dynamic of complex technical and natural systems is usually nonlinear<br />
and therefore the prediction of the dynamic behaviour of such systems can<br />
only be done if a powerful simulation and investigation tool is available. AnT<br />
17
18<br />
Staff<br />
(www.Ant4669.de) is such a tool that supports a broad spectrum of system classes<br />
and analysis methods.<br />
Staff of the Department:<br />
(from left ) (front) T. Dipper, B. Eckstein, K. Häussermann, U. Käppeler, A. Koch, P. Levi, U. Gräter,<br />
M. Schanz, S. Kernbach, B. Schießle, D. Sutantyo, E. Meister<br />
(back) B. Schenke, V. Avrutin, D. Di Marco, F. Schlachter, S. Popesku, O. Zweigle, O. Kernbach<br />
(staff not shown on the photo) F. Aichele, H. Rajaie, A. Tamke
Research Projects<br />
www.ipvs.uni-stuttgart.de/abteilungen/bv/forschung/projekte<br />
1. RFC-<strong>Stuttgart</strong><br />
(formerly known as CoPS - Cooperative Soccer Playing Robots<br />
<strong>Stuttgart</strong>)<br />
Personnel: O. Zweigle, H. Rajaie, K. Häussermann, U. Käppeler, A. Tamke, A.<br />
Koch, B. Eckstein, M. Schanz<br />
F<strong>und</strong>ed by: <strong>Universität</strong> <strong>Stuttgart</strong><br />
Duration: since 1.7.2001<br />
RoboCup is an international joint initiative to promote artifi cial intelligence (AI), robotics,<br />
and related fi elds. In 2009 the robots of our Team 1.RFC <strong>Stuttgart</strong> were improved<br />
both in hard- and soft ware. Th us an accurate passing system has been developed<br />
which was able to adapt the passing behavior during the game. Furthermore we<br />
improved the learning algorithms to adapt the tactics and strategic behavior in realtime<br />
as well was the motor controller to improve the driving behavior has been optimized.<br />
RoboEarth - Connecting Robots Worldwide<br />
Personnel: A. Koch, K. Häussermann, B. Schießle, D. DiMarco<br />
F<strong>und</strong>ed by: European Community (EC)<br />
Duration: 1.12.2009 - 30.11.2013<br />
RoboEarth is a world-wide platform which enables robots to exchange position and<br />
map information as well as task-related, hardware-independent action recipes. In 2009<br />
we developed fi rst approaches to create the general architecture. Furthermore we discussed<br />
and evaluated some defi nitions and agreements of the centralized and platform<br />
independent web-based database system, which will be able to manage diff erent robotic<br />
relevant data.<br />
SFB 627 - Project E3 - Distributed Situation Recognition Based on<br />
Evaluated Context Information<br />
Personnel: K. Häussermann<br />
F<strong>und</strong>ed by: Deutsche Forschungsgemeinschaft (DFG)<br />
Duration: 1.1.2007 - 31.12.2010<br />
Th e objectives of sub-project E3 is to develop a distributed application specifi c situation<br />
recognition based on evaluated context information. In the year 2009 a new approach<br />
has been developed to calculate and adapt the situation templates, based on<br />
the quality of the context sources. Furthermore a method to detect and solve inconsistencies<br />
using a so called meta-template has been developed and evaluated accordingly.<br />
19
20<br />
SFB 627 – Project C3 – Sensor-Supported, Context-Based Consistency<br />
Analysis<br />
Personnel: U. Käppeler<br />
F<strong>und</strong>ed by: Deutsche Forschungsgemeinschaft (DFG)<br />
Duration: 1.1.2003 – 31.12.2010<br />
One of the main objectives of the subproject C3 is to develop concepts for the evaluation<br />
of the quality of sensor data, in order to integrate indistinct measurements and<br />
inaccurate results into distributed environment models. In the year 2009 a method to<br />
derive higher level context or situation information combined with probabilities from<br />
inaccurate sensor data using a multinomial logistic regression has been developed and<br />
evaluated successfully.<br />
Replicator – Robotic Evolutionary Self-Programming and Self-Assembling<br />
Organisms<br />
Personnel: S. Kernbach, E. Meister, S. Popesku<br />
F<strong>und</strong>ed by: European Community (EC)<br />
Duration: 1.3.2008 - 28.2.2013<br />
Th is project as well as the project SYMBRION are both EC-f<strong>und</strong>ed projects. Th ere<br />
are big synergies between both projects however some essential points diff er. In REP-<br />
LICATOR the main goal is to build a sensor- and communication-rich platform that<br />
are capable of self-aggregation into large-scale artifi cial organisms. Flexible locomotion<br />
and adaptation are the main focuses and should be achieved autonomously without<br />
human intervention. A big challenge is to develop a highly adaptive hardware in<br />
cooperation with highly adaptive soft ware.<br />
Symbrion – Symbiotic Evolutionary Robot Organisms<br />
Personnel: S. Kernbach, O. Kernbach, F. Schlachter<br />
F<strong>und</strong>ed by: European Community (EC)<br />
Duration: 1.2.2008 - 31.1.2013<br />
Th e main focus of this project is to investigate novel bio-inspired approaches for symbiotic<br />
multi-robot organisms. Principles of evolution and adaptation in large-scale artifi<br />
cial organisms play the main role of the research focus. Based on experiences and<br />
knowledge from former projects such as I-SWARM or MiCRON, which both dealt<br />
with emergent swarm behavior, as a next step aggregation into organisms that are physically<br />
connected have been developed. As a result at the end of this project, it is expected<br />
that such artifi cial organisms will be capable of self-confi guration, self-healing,<br />
self-adaptation or self-protection and hence human intervention should be avoided.<br />
Such systems can be used for rescue tasks in hazardous environments where presence
of humans either not possible or perilous. Th e SYMBRION project involves several<br />
benefi ciaries from the REPLICATOR projects and vice versa.<br />
ANGELS – Anguilliform Robot with Electric Sense<br />
Personnel: S. Kernbach, D. Sutantyo<br />
F<strong>und</strong>ed by: European Community (EC)<br />
Duration: 1.2.2009 - 31.1.2012<br />
Th e aim of the ANGELS project is to design and build a prototype of a reconfi gurable<br />
swimming robot able to split into smaller agents (and vice-versa). Compared to<br />
the present existing solutions for Autonomous Underwater Vehicles (AUV) exploration,<br />
ANGELS fi rst exploits the advantages of using a bio-inspired sense that has not<br />
been extensively explored by robotics until now: the electric sense. On this point, a<br />
new promising sensor technology is introduced, that can be applied in environmental<br />
exploration and inspection where vision cannot be used. Moreover, in order to investigate<br />
the <strong>und</strong>erwater collective behavior, a test platform robot is also prepared. Th e<br />
ANGELS test platform uses a small size <strong>und</strong>erwater swimming toy as the mechanical<br />
platform and blue light sensor for communication among robots.<br />
Selected Research Projects<br />
www.ipvs.uni-stuttgart.de/abteilungen/bv/forschung/projekte<br />
1. RFC-<strong>Stuttgart</strong> (formerly known as CoPS - Cooperative Soccer<br />
Playing Robots <strong>Stuttgart</strong>)<br />
RoboCup is an international joint initiative to promote artifi cial intelligence (AI), robotics,<br />
and related fi elds. It is an attempt to boost robotic research and artifi cial intelligence<br />
by providing a standard problem where a wide range of technologies can be integrated<br />
and examined.<br />
Th e 1. RFC-<strong>Stuttgart</strong> (formerly<br />
known as the CoPS<br />
team) is a middle size team<br />
of the Department of Image<br />
Understanding of the IPVS.<br />
In 2009 the new redesigned<br />
robot hardware with a four<br />
wheeled omnidirectional<br />
drive was enhanced in several<br />
parts of hardware and<br />
soft ware. Th e hardware has<br />
21
22<br />
been enhanced by new constructed wheels and by the development of a new digital<br />
compass to improve the localization of the robot. Furthermore the adjustable electromagnetic<br />
kicker device was boosted to allow stronger ball shoots.<br />
One of the major soft ware topics in 2009 were the development of an accurate<br />
passing system which is able to adapt the passing behavior during the game, the improvement<br />
of learning algorithms to adapt the tactics and strategic behavior in realtime<br />
and also the optimization of the motor controller to improve the driving behavior.<br />
Furthermore to analyze a game in real-time an autonomous camera-agent was<br />
developed, which is driven only by communicated robot data in combination of a<br />
highly sophisticated 3D Visualization.<br />
Th e team won the German Open 2009 in Hanover and even during the world<br />
championship of Robocup in Graz (Austria), the 1.RFC-<strong>Stuttgart</strong> became new world<br />
champion 2009 in the the middle-size league. Furthermore the team achieved the 2nd<br />
place at the so-called “technical challenge” and a further 1st place at the “scientifi c<br />
challenge”.<br />
RoboEarth - Connecting Robots Worldwide<br />
RoboEarth is a world-wide platform which robots can use to exchange position and<br />
map information as well as task-related, hardware-independent action recipes. Th is will<br />
enable manufacturers worldwide to break down their costs and eff orts for reproducing<br />
soft ware algorithms for robot behavior over and over again. Th e RoboEarth framework<br />
can store all relevant data from algorithms to complex behavior descriptions that<br />
allows robots to act autonomously in an unknown, unspecifi ed environment. In 2009<br />
some literature surveys and fi rst approaches have been developed to create the general<br />
architecture of the so called recognition-/labeling-component, which will serve as the<br />
basic interface between the robot platforms and RoboEarth. Furthermore fi rst defi -<br />
nitions and agreements of<br />
the architecture of the centralized<br />
and platform independent<br />
web-based database<br />
system, which will<br />
be able to manage diff erent<br />
robotic relevant data,<br />
has been discussed and<br />
evaluated.
2.3 Parallel Systems<br />
<strong>Universität</strong> <strong>Stuttgart</strong> - IPVS<br />
<strong>Parallele</strong> <strong>Systeme</strong><br />
<strong>Universität</strong>sstraße 38<br />
70569 <strong>Stuttgart</strong><br />
Germany<br />
Research Focus<br />
Prof. Dr.-Ing. Sven Simon<br />
� Sven.Simon@ipvs.uni-stuttgart.de<br />
� +49-711-7816450<br />
� +49-711-7816250<br />
Th e research activities of the department Parallel Systems involve the analysis and performance<br />
optimization of hardware/soft ware systems with the focus on the interaction<br />
of the diff erent levels of the system design. Th is includes the physical design level<br />
of the hardware platform, the architectural level of the devices and the algorithmic<br />
level of the application. Prototypes are built to demonstrate the effi ciency of the approach.<br />
In order to obtain very compact high performance systems the resulting hardware/soft<br />
ware systems are domain specifi c with a focus on the following research areas:<br />
▶ Parallel Real-Time Image and Video Processing<br />
In many industrial scenarios like the monitoring of manufacturing processes, the<br />
production process is controlled using image and video processing. Today’s image<br />
sensors enable resolutions well above one megapixel with frame rates exceeding<br />
several h<strong>und</strong>red fps which are well suited for monitoring even fast processes.<br />
While the raw video material can be obtained easily, its analysis using CPU-based<br />
hardware and algorithms implemented in soft ware is very time-consuming due<br />
to the huge amount of data involved. In the case of several h<strong>und</strong>red of high-resolution<br />
images per second real-time processing can be achieved by reconfi gurable<br />
hardware in conjunction with highly-parallelized algorithms. In order to apply<br />
this to process control, the latency of the algorithms mapped on the specifi c hardware<br />
platform is an important issue.<br />
▶ Hardware Accelerators for Scientifi c Computing<br />
In order to reduce the computation time in scientifi c computing, typically PC<br />
clusters or supercomputers are used. Usually, these computing resources have to<br />
be shared with other users due to cost reasons such that a certain waiting time<br />
or latency has to be accepted by the user which is the contrary of the intended<br />
acceleration. As an alternative, hardware accelerators based on GPGPUs or reconfi<br />
gurable hardware (FPGAs) can be used in personal computers such that the<br />
23
24<br />
computing resources are available when required. Th e programming model and<br />
architecture of hardware accelerators diff er very much from classical parallel computers<br />
such that there is a high demand of research to investigate the achievable<br />
performance in certain application domains.<br />
▶ Hardware Design Techniques for Multi-Gigabit Systems<br />
Th e computational effi ciency of highly parallel hardware architectures requires<br />
multi-gigabit data transfer between diff erent hardware components. Although<br />
system-on-the-chip solutions integrate components with suffi cient bandwidth,<br />
the off -chip data transfer bandwidth to external components has become a critical<br />
issue over the years. In order to implement systems with multi-gigabit interconnects<br />
in the range of 10Gbits/s to 100Gbit/s or above in conjunction with<br />
complex IC devices new design techniques for the integration of heterogeneous<br />
components have to be developed. Particular eff ort is made to enable accurate and<br />
fast simulation of the physical level of these systems by appropriate measurement<br />
and modeling. Important aspects are signal propagation, power distribution, variation<br />
of manufacturing parameters and signal integrity.<br />
Staff<br />
Staff of the<br />
Department:<br />
(from left )<br />
S. Simon,<br />
M. Shaikh,<br />
L. Rockstroh,<br />
M. Wróblewski,<br />
J. Hillebrand,<br />
T. Bringewat,<br />
U. Gräter,<br />
S. Wahl,<br />
W. Li<br />
(staff not shown on the<br />
photo)<br />
M. Grégoire,<br />
C. Qiu,<br />
Z. Wang
Research Projects<br />
www.ipvs.uni-stuttgart.de/abteilungen/pas/forschung/projekte<br />
High Performance Hardware Platforms<br />
Personnel: M. Wróblewski, J. Hillebrand<br />
F<strong>und</strong>ed by: University of <strong>Stuttgart</strong><br />
Duration: Unlimited<br />
Th e goal of the work is the design of high performance hardware platforms based on<br />
reconfi gurable devices, general purpose processors and graphics processing units for<br />
specifi c application domains.<br />
Numerical accuracy analysis in scientifi c computing<br />
Personnel: W. Li<br />
F<strong>und</strong>ed by: Simtech, Deutsche Forschungsgemeinschaft (DFG)<br />
Duration: 1.8.2008 – 30.7.2011<br />
Numerical accuracy analysis is a general concern in scientifi c simulation applications<br />
to reduce the risk of misinterpretation of inaccurate results due to fi nite precision<br />
computation and ro<strong>und</strong>ing error propagation. Th e research in our project provides an<br />
eff ective and effi cient numerical accuracy analysis method as well as an easy to use numerical<br />
accuracy debugging tool.<br />
Enhancement of the Effective I/O-Bandwidth of High Performance<br />
Systems<br />
Personnel: S. Wahl, Z. Wang<br />
F<strong>und</strong>ed by: <strong>Universität</strong> <strong>Stuttgart</strong><br />
Duration: since 1.9.2007<br />
Th e increasing diff erence between the performance of processor cores and the I/Obandwidth<br />
leads to limitations which can be overcome by improved eff ective I/Odata.<br />
Th is effi ciency enhancement is especially applicable in reconfi gurable hardware<br />
systems on the bit-level but can also be extended to the general purpose processor domain.<br />
A Sensor Data Platform for Monitoring Production Processes<br />
Personnel: L. Rockstroh<br />
F<strong>und</strong>ed by: Graduate School advanced Manufacturing Engineering (GSaME)<br />
Duration: 1.5.2008 – 30.4.2012<br />
Th e focus of this project is a scalable hardware platform for sensor data processing in a<br />
production environment with emphasis on real time constraints and robustness.<br />
25
26<br />
A Wireless Transmitter Concept Based on Sigma-Delta-Synthesizers<br />
F<strong>und</strong>ed by: B<strong>und</strong>esministerium <strong>für</strong> Bildung <strong>und</strong> Forschung (BMBF)<br />
Personnel: K. Merkle – external<br />
Duration: 30.9.2007 – 30.11.2010<br />
In this external research project a circuit concept for a wireless transmitter is investigated<br />
which replaces the A/D-converter of a wireless transmitter in order to simplify<br />
the hardware concept.<br />
Acceleration of Algorithms for In-Process Measurement Techniques<br />
of Nanostructured Surfaces<br />
Personnel: M. Shaikh, M. Grégoire<br />
F<strong>und</strong>ed by: Deutsche Forschungsgemeinschaft (DFG)<br />
Duration: 15.6.2009 – 14.6.2011<br />
In this project the algorithmic part of measurement techniques with respect to acceleration<br />
is examined which can be applied to nanostructured surfaces. Th e work is<br />
carried out in collaboration with the <strong>Institut</strong>e of Measurement Techniques of Bremen<br />
University.<br />
Acceleration of Simulations of Electromagnetic Wave Propagation<br />
in Biomedical Applications<br />
Personnel: M. Grégoire, T. Bringewat<br />
F<strong>und</strong>ed by: Baden-Württemberg Stiftung<br />
Duration: 15.6.2009 – 14.6.2011<br />
In collaboration with the “<strong>Institut</strong> <strong>für</strong> Lasertechnologien in der Medizin <strong>und</strong><br />
Meßtechnik” at the University of Ulm the accelerated simulation of electromagnetic<br />
wave propagation in biological cells are examined based on the classical fi nite diff erence<br />
time domain (FDTD) approach and the discrete dipole approximation (DDA)<br />
approach.<br />
Modeling and Analysis of High Speed Multi-Gigabit Interconnects<br />
Personnel: J. Hillebrand<br />
F<strong>und</strong>ed by: <strong>Universität</strong> <strong>Stuttgart</strong><br />
Duration: since 10.1.2009<br />
Th e project aims to derive design and measurement methodologies for board-level interconnects<br />
of future multi-gigabit data transmission standards. Th e focus lies on experimental<br />
investigation, analysis and subsequent modeling of physical eff ects relevant<br />
to the design of printed circuit boards carrying the signals of these high speed communication<br />
standards.
Selected Research Projects<br />
A Sensor Data Platform for Monitoring of Production Processes<br />
Th ermal spray processes are particle-based manufacturing techniques that deposit<br />
coating material on the surface of work pieces. Th e feedstock material used for thermal<br />
spraying consists of particles with diameters in between 0.1 and 50 micrometers that<br />
are heated to a molten state and sprayed on the work piece with velocities of up to 900<br />
meters per second. Both, coating quality and the amount of overspray (wasted feedstock<br />
material) are correlated to particle characteristics such as distribution and velocities<br />
of the particles. Th us measuring the distribution and velocities of the particles is<br />
essential in order to predict coating quality and overspray.<br />
Continuous Particle Image Velocimetry (CPIV) is a method for measuring particle<br />
distribution and velocity. It is based on motion blurring of particles due to the<br />
exposure time of the imaging system and the velocity of the particles. Th e measurement<br />
setup of Continuous Particle Image Velocimetry (CPIV) is the least complex<br />
setup compared to other image-based methods such as Classical Particle Image Velocimetry<br />
and Particle Tracking Velocimetry which predestines Continuous PIV for<br />
industrial applications. Current soft ware-based implementations of Continuous PIV<br />
achieve aro<strong>und</strong> a few measurement samples per second due to image analysis computations<br />
with complex data dependencies. However, online process monitoring and control<br />
loops require one order of magnitude higher measurement sample rates which can<br />
be achieved by using a highly parallelized and optimized hardware accelerator architecture<br />
implemented on a Field Programmable Gate Array (FPGA).<br />
Implementing a hardware-accelerated version of Continuous PIV requires developing<br />
an effi cient algorithm in order to utilize the parallel resources of the FPGA. Th e<br />
algorithm consists of the following steps:<br />
▶ Step 1: Segmentation of the Image<br />
▶ Step 2: Classifi cation and Particle Velocimetry calculations<br />
(a) original image (b) based on maximum intensity<br />
(c) based on mean intensity (d) combined (b and c)<br />
Segmentation of Continuous PIV<br />
images<br />
27
28<br />
Th e fi rst step is to determine for each pixel whether it belongs to the foregro<strong>und</strong>, which<br />
is considered to consist of objects, or to the backgro<strong>und</strong> of the image. Values of pixels<br />
that are segmented into the foregro<strong>und</strong> are maintained as these values are required for<br />
the classifi cation step while pixels of the backgro<strong>und</strong> are set to zero. Th e decision criterion<br />
for the segmentation is the intensity of the pixel in dependence to the mean intensity<br />
and the largest intensity of neighbouring pixels. Th is combination enables the<br />
detection of dim particles in noisy environments such as shown in the fi gure above as<br />
well as the detection of bright particles.<br />
During classifi cation in step 2, each cluster of non-zero pixels is considered as an<br />
object and the extent of these objects is determined by analyzing for each pixel of the<br />
cluster, how it contributes to the length and width of the object.<br />
Th e fi gure below illustrates the projections of particle objects. Long cylindrical<br />
objects without signifi cant change of width and intensity are considered particle projections.<br />
Based on start and end position of these projections, length and heading are<br />
determined. Finally, the velocity is calculated by taking the edge length of the image<br />
sensor and the magnifi cation factor into account.<br />
Particle objects occurring<br />
in continuous PIV images<br />
(a) single particle (b) overlapped (width)<br />
(c) overlapped (intensity) (d) fractions of a particle<br />
Accelerating the Continuous PIV algorithm is achieved by an effi cient hardwarearchitecture<br />
that is implemented on a FPGA. Th e architecture consists of two processing<br />
units that are arranged in a pipeline structure and each of them is related to<br />
one step of the algorithm. Aft er an initialization stage, both units work concurrently<br />
which results in a speedup compared to purely sequential processing. In addition, each<br />
of them is adapted specifi cally to its task and contains several processing elements that<br />
are arranged in a structure of parallel pipelines.<br />
Numerical accuracy analysis in scientifi c computing<br />
Th e focus of this project is to investigate the impact of a reduced fl oating point mantissa<br />
word length on performance enhancements of fl oating point algorithms like sparse<br />
matrix solvers. Usually, signifi cant hardware savings can be obtained on hybrid high<br />
performance systems (GPUs, FPGAs) by the reduction of the mantissa word length<br />
for arithmetic operations which can be used for further parallelization and speed up.
However, for any given algorithm, the infl uence of the word length reduction on the<br />
accuracy of the fi nal result is not obvious. Even worse, it is open whether the classical<br />
double precision mantissa word length is suffi cient for the accuracy of widely used<br />
simulation soft ware packages, taken into account that these packages contain millions<br />
of lines of code implemented over years by a large group of persons.<br />
In order to verify the dependency of numerical accuracy on the mantissa wordlength<br />
a method based on Discrete Stochastic Arithmetic is used. Th e fi nite word<br />
length eff ects in the arithmetic logic are randomized by changing the behavior (ro<strong>und</strong>ing,<br />
truncation, etc.) for every operation executed, thus yielding each time a diff erent<br />
result R, and therefore a diff erent deviation ε=R-r from the mathematically exact result<br />
r. Th is enables obtaining of statistically signifi cant results based on only a few repeated<br />
executions of a particular algorithm.<br />
�����<br />
P���<br />
Th e method has been shown to work well with several data fl ow examples (e.g.<br />
matrix-multiplication) as well as some widely used soft ware packages (e.g. LAPACK,<br />
FFTW, R) involving billions of fl oating point operations. In all these experiments, the<br />
used method showed its eff ectiveness and reliability. Th e method has also been applied<br />
to more complex programs like MOLPRO, a quantum chemistry package with several<br />
100.000 lines of source code. One focus of the investigations is the hardware acceleration<br />
of the proposed method.<br />
�<br />
Random ro<strong>und</strong>ing<br />
analysis of the MOLPRO<br />
package. Probability P<br />
(yellow) of obtaining a<br />
particular deviation ε from<br />
mathematically exact result<br />
vs. the value of the deviation<br />
ε. Th e interpolated curve<br />
(red) approaches the Gauss<br />
distribution (blue)<br />
29
30<br />
2.4 Simulation of Large Systems<br />
<strong>Universität</strong> <strong>Stuttgart</strong> - IPVS<br />
Simulation großer <strong>Systeme</strong><br />
<strong>Universität</strong>sstraße 38<br />
70569 <strong>Stuttgart</strong><br />
Germany<br />
Research Focus<br />
Prof. Dr. rer. nat. Marc Alexander Schweitzer<br />
� Marc.Alexander.Schweitzer<br />
@ipvs.uni-stuttgart.de<br />
� +49-711-7816465<br />
� +49-711-7816248<br />
Prof. Marc Alexander Schweitzer, from the <strong>Institut</strong>e for Numerical Simulation at University<br />
of Bonn, accepted in early 2010 the position of head of chair and started in<br />
May. Prof. Schweitzer’s research focus has been on mesh-free methods for the solution<br />
of partial diff erential equations, multilevel and algebraic multigrid methods, as well as<br />
on parallelization and effi cient algorithms for simulation problems.<br />
Jun.-Prof. Michael Bader, formerly at the Department of Informatics at Technische<br />
<strong>Universität</strong> München, started as a SimTech Juniorprofessor in December 2009.<br />
He leads a young researcher’s group on Simulation Soft ware Engineering. His research<br />
focus is on effi cient algorithms in numerical simulation, especially on effi cient approaches<br />
for parallel adaptive mesh refi nement.<br />
Research Projects<br />
www.ipvs.uni-stuttgart.de/abteilungen/sgs/forschung/projekte<br />
SRC: SimTech - Recursively Structured Triangular and Tetrahedral<br />
Grids for Dynamically Adaptive Simulations in DUNE<br />
Personnel: M. Bader, K. Rahnema<br />
F<strong>und</strong>ed by: Deutsche Forschungsgemeinschaft (DFG)<br />
Duration: 1.2.2010 – 31.1.2013<br />
Th e project is embedded in the Cluster of Excellence “Simulation Technology” (Sim-<br />
Tech), and integrated in SimTech’s research network on multi-physics simulation. Th e<br />
goal is to integrate an existing approach for dynamically adaptive grid generation and<br />
grid based numerical simulation into the problem solving environment DUNE. Demonstrator<br />
applications will be chosen from porous media fl ow (in the context of CO2<br />
sequestration) and oceanic wave propagation (Tsunami simulation).
2.5 Distributed Systems<br />
<strong>Universität</strong> <strong>Stuttgart</strong> - IPVS<br />
<strong>Verteilte</strong> <strong>Systeme</strong><br />
<strong>Universität</strong>sstraße 38<br />
70569 <strong>Stuttgart</strong><br />
Germany<br />
Research Focus<br />
Prof. Dr. rer. nat. Dr. h.c. Kurt Rothermel<br />
� Kurt.Rothermel@ipvs.uni-stuttgart.de<br />
� +49-711-7816434<br />
� +49-711-7816424<br />
A wide spectrum of distributed applications involves a potentially large number of<br />
mobile devices, which might substantially diff er in performance and degree of mobility.<br />
Oft en those applications are based on various network technologies, including<br />
infrastructure-based networks, mobile ad hoc networks and sensor networks. Further,<br />
many of those applications exploit context information to dynamically adapt their behavior<br />
according to the current user situation. Th e research activities of the distributed<br />
systems group focus on system-level aspects of those applications, in particular methods<br />
for distributed data management and communication are investigated. Currently,<br />
our research addresses the following areas:<br />
▶ Context-aware systems<br />
utilize information about the physical world to adapt automatically to the context<br />
of users. Here, we investigate mechanisms for realizing large-scale distributed context<br />
models and mechanisms for supporting mobile context-aware applications.<br />
Th is research is done within the Collaborative Research Center Nexus (SFB 627).<br />
▶ Adaptive Communication Systems<br />
Distributed applications deployed on large scale need to cope with a high degree<br />
of network dynamicity because of changes to the set of participating processes,<br />
failures or mobility of processes. Th e focus of our research is to support robust<br />
and reliable applications by devising communications abstractions with the ability<br />
to dynamically adapt according to the availability of resources.<br />
▶ Self-organizing soft ware systems<br />
A self-organizing soft ware system is able to adapt its structure to changing environmental<br />
conditions to optimize its operation. Th is is especially important for<br />
systems that do not allow manual maintenance due to their scale and dynamics.<br />
Here, we investigate new mechanisms and concepts for the self-organization in<br />
such systems.<br />
31
32<br />
▶ Sensor Networks<br />
Wireless sensor networks have been proposed as a way to unobtrusively gather<br />
real-world data. Typically, they consist of small networked devices equipped with<br />
sensors. Th e focus of our research lies on the soft ware architecture and algorithms<br />
needed to deal with the increasing complexity of those systems.<br />
▶ Grid Computing<br />
is a discipline of high performance computing which deals with federating resources<br />
that belong to diff erent administrative domains and are oft en globally distributed.<br />
In our research we investigate on how to employ Peer-to-Peer principles<br />
to cope with the high degree of volatility and heterogeneity of resources pooled<br />
in Grid environments.<br />
In the year 2009, the Distributed Systems group together with the Collaborative Research<br />
Center Nexus organized the First International Workshop on Quality of Context<br />
(QuaCon 2009). QuaCon 2009 was the fi rst scientifi c event that specifi cally focused<br />
on the diff erent aspects of quality of context information. It aimed to bring<br />
together researchers from various fi elds to discuss diff erent aspects of context quality<br />
and to make a consolidated contribution towards an integrated way of treating this<br />
topic. Th e workshop was well received by more than 100 researches from Europe, Asia,<br />
and USA.<br />
Moreover, members of the Distributed Systems group presented their work at<br />
various prestigious scientifi c events, in particular, Eighth IEEE International Conference<br />
on Pervasive Computing and Communications (PerCom 2010), Sixth International<br />
Conference on Mobile and Ubiquitous Systems (MobiQuitous 2009), and International<br />
Conference on Mobile Data Management (MDM 2009).
Staff<br />
Staff of the Department:<br />
(from front) (left row) A. Grau, H. Weinschrott, W. Blochinger, K. Rothermel, S. Schuhmann, A. Benzing,<br />
B. Koldehofe, L. Geiger, M. Wernke, Ch. Hiesinger<br />
(middle row) G. Koch, B. Ottenwälder, M. Brodbeck, S. Rizou, A. Roesler, S. Föll, B. Schilling, M. Völz<br />
(right row) A. Tariq, B. Hameed, P. Skvorzov, S. Schnitzer, D. Fischer, H. Wolf, K. Herrmann, F. Dürr<br />
(staff not shown on the photo) R. Lange, D. Philipp, P. Schlegel, S. Schulz<br />
Research Projects<br />
www.ipvs.uni-stuttgart.de/abteilungen/vs/forschung/projekte<br />
ALLOW - Adaptable Pervasive Flows<br />
Personnel: K. Herrmann, H. Wolf, C. Hiesinger, S. Föll<br />
F<strong>und</strong>ed by: European Community (EC)<br />
Duration: 1.2.2008 – 31.1.2011<br />
ALLOW is a European research project f<strong>und</strong>ed <strong>und</strong>er the 7th Framework Programme.<br />
Its objective is to develop a new programming paradigm for human-oriented adaptable<br />
pervasive applications. Th e core concepts of the project are Adaptable Pervasive Flows<br />
(APFs). APFs are employed for modeling the activities of the users involved in a per-<br />
33
34<br />
vasive application. Th is model is then used for adapting the environment to the user<br />
and his current needs.<br />
Parallel SAT Solving Methods<br />
Personnel: W. Blochinger, S. Schulz<br />
F<strong>und</strong>ed by: Deutsche Forschungsgemeinschaft (DFG)<br />
Duration: 1.7.2006 - 30.6.2010<br />
SAT solving checks whether there exists a satisfying variable assignment for a given<br />
Boolean formula. It has become an essential tool in many application domains, e.g.,<br />
electronic design automation, soft ware verifi cation, or confi guration of complex products.<br />
Th e objective of this project is to signifi cantly speed up SAT solving methods by<br />
parallel processing, extending the range of current and paving the way for future applications.<br />
In the second f<strong>und</strong>ing period, we focus on methods and systems for exploiting<br />
the distributed, highly volatile, and heterogeneous resources of Peer-to-Peer Desktop<br />
Grids for parallel SAT solving.<br />
Global Sensor Grid<br />
Personnel: B. Koldehofe, A. Benzing<br />
F<strong>und</strong>ed by: Cluster of Excellence Simulation Technology<br />
Duration: 1.5.2008 - 30.4.2011<br />
Th e goal of the project is to provide gapless real-time information to simulations on a<br />
global scale. To fi ll the gaps in existing sensor network deployments, so-called diagnostic<br />
simulations are distributed in the Global Sensor Grid. Flexible access to sensor data<br />
is provided by supporting moving queries that can be dynamically relocated to follow<br />
the observed phenomenon. Th e project is part of and f<strong>und</strong>ed by the Cluster of Excellence<br />
Simulation Technology.<br />
SpoVNet 2.0<br />
Personnel: B. Koldehofe, G. Koch, A. Tariq<br />
F<strong>und</strong>ed by: Baden-Württemberg Stiftung<br />
Duration: 1.1.2010 - 31.12.2011<br />
SpoVNet 2.0 is a transfer project for the precursory SpoVNet project where a methodology<br />
and an infrastructure were developed to deploy application-oriented services in<br />
heterogeneous communication networks in a spontaneous and adaptive way. In addition,<br />
SpoVNet 2.0 investigates the modularization and intelligent composition of the<br />
services and their integration into the Internet of Th ings and adaptation to network<br />
virtualization.
Reliability in Distributed Complex Event Processing<br />
Personnel: M. Völz<br />
F<strong>und</strong>ed by: <strong>Universität</strong> <strong>Stuttgart</strong><br />
Duration: 15.1.2009 - 31.12.2010<br />
While distributed Complex Event Processing systems can meet the scalability requirements<br />
of modern applications, any requirements concerning reliability remain unaddressed<br />
so far. Th is project aims to provide distributed CEP systems with the needed<br />
reliability, e.g. by masking failures through replication or reactive failure compensation.<br />
Furthermore, performance aspects such as effi cient placement of replicated components<br />
are considered.<br />
DHEP – Distributed Heterogeneous Event Processing<br />
Personnel: B. Koldehofe, B. Schilling<br />
F<strong>und</strong>ed by: IBM<br />
Duration: 1.3.2008 – 28.2.2010<br />
Th e project “Distributed Heterogeneous Event Processing” explores new ways in order<br />
to increase the scalability and availability of existing complex event processing technology<br />
by establishing a distributed event correlation network. Th erefore, a framework<br />
is developed that comprises heterogeneous correlation engines to enable effi cient<br />
event processing. Th e focus lies in providing extensive functionality by the introduction<br />
of a highly expressive description language and adaptive rule placement strategies<br />
that account for the heterogeneity of the <strong>und</strong>erlying network.<br />
Complex Query Optimization in P2P Overlay Networks<br />
Personnel: F. A. Memon<br />
F<strong>und</strong>ed by: <strong>Universität</strong> <strong>Stuttgart</strong><br />
Duration: since 1.12.2006<br />
Th is project deals with the optimization of complex queries in P2P overlay networks,<br />
with main focus on multi-attribute range queries. Given a set of queries that have been<br />
previously monitored in a P2P overlay network, the goal is to optimize the search index<br />
for these queries. Such an optimization leads to signifi cant improvement in overall<br />
system performance in terms of scalability.<br />
UP2P - Uniform Peer-to-Peer Protocols for Heterogenous Networks<br />
Personnel: B. Koldehofe<br />
F<strong>und</strong>ed by: Deutscher Akademischer Austauschdienst (DAAD)<br />
Duration: 1.1.2009 - 31.12.2010<br />
Th e project on uniform peer-to-peer computing (UP2P) is a research cooperation between<br />
the IPVS and the University of Lisbon to establish and intensify cooperation<br />
35
36<br />
between young researchers and Master Students in Germany and Portugal. Th e goal is<br />
to identify overlay structures that can be deployed at low cost over heterogeneous networks,<br />
for instance over ad-hoc and infrastructure networks.<br />
AWARE - Platform for Autonomous self-deploying and operation of<br />
Wireless sensor-actuator networks cooperating with AeRial objects<br />
Personnel: K. Herrmann, M. Gonzalo<br />
F<strong>und</strong>ed by: European Community (EC)<br />
Duration: 5.6.2006 - 4.6.2009<br />
AWARE was a European research project f<strong>und</strong>ed <strong>und</strong>er the 6th Framework Programme.<br />
Its focus was to develop a platform for enabling the cooperation of autonomous<br />
aerial vehicles with gro<strong>und</strong> wireless sensor-actuator networks, comprising static<br />
and mobile nodes. Especially in disaster relief scenarios, the results of the AWARE<br />
platform will permit operations in areas that are diffi cult to access and that do not<br />
provide a communication infrastructure. Th is is achieved by combining diverse sensor<br />
systems. Th e distributed system research group was involved in the development of a<br />
respective middleware platform. Th e project was very successfully completed in August<br />
2009. Th e consortium has won the second prize in the “Robotics 2010 Awards”<br />
with the UAV platform that was developed during the project.<br />
NET – Network Emulation Testbed<br />
Personnel: K. Herrmann, A. Grau<br />
F<strong>und</strong>ed by: Deutsche Forschungsgemeinschaft (DFG)<br />
Duration: 1.4.2001 - 31.8.2011<br />
Th e NET project is building an emulation testbed for running evaluations of real soft -<br />
ware systems. New concepts of node and time virtualization are used on a computer<br />
cluster to achieve a versatile emulation environment that uses the cluster’s resources<br />
optimally. In conjunction with effi cient placement strategies for virtual nodes, the emulation<br />
testbed allows for running large-scale network experiments in minimal time.<br />
Optimized Information Discovery<br />
Personnel: F. Dürr, F. Memon<br />
F<strong>und</strong>ed by: <strong>Universität</strong> <strong>Stuttgart</strong><br />
Duration: 1.12.2006 – 30.11.2010<br />
Th e optimized information discovery (OID) project deals with optimizing the overall<br />
system performance of multi-attribute range (MAR) queries in distributed-hash-tablebased<br />
information discovery systems. Th e goal of this project is to create and dynamically<br />
adapt a set of indices over DHTs. Th e adaptation process is based upon the history<br />
of MAR queries.
GSaME – Project E5 – Novel Communication Architecture for the<br />
Smart Factory<br />
Personnel: F. Dürr, B. Hameed<br />
F<strong>und</strong>ed by: Graduate School advanced Manufacturing Engineering (GSaME)<br />
Duration: 1.8.2008 – 31.7.2012<br />
Project E5 investigates consistency issues in product monitoring inside a smart realtime<br />
factory. As a fi rst step, a consistency stack has been developed that conceptualizes<br />
the diff erent consistency issues into separate layers. In addition to this, an algorithm<br />
has been designed to provide probabilistic guarantees about the observation of objects<br />
and the sequences in which these objects are moving through the assembly lines.<br />
Current work is now focused on developing adaptive strategies to dynamically adjust<br />
the probabilities for correct readings of RFID devices depending upon their historical<br />
performance.<br />
SFB 627 – Project A2 – Context-based Communication<br />
Personnel: F. Dürr, L. Geiger<br />
F<strong>und</strong>ed by: Deutsche Forschungsgemeinschaft (DFG)<br />
Duration: 1.1.2007 – 31.12.2010<br />
Project A2 researches protocols to effi ciently disseminate messages addressed to users<br />
in a certain context instead of directly addressing individuals. Since an effi cient dissemination<br />
requires up-to-date information about users, we currently focus our investigations<br />
on aggregation methods for user contexts to lower the update load.<br />
SFB 627 – Project B3 – Hybrid Model Management<br />
Personnel: F. Dürr, H. Weinschrott<br />
F<strong>und</strong>ed by: Deutsche Forschungsgemeinschaft (DFG)<br />
Duration: 1.1.2003 – 31.12.2010<br />
Th e Project B3 of the SFB 627 investigates mechanisms to acquire and manage dynamic<br />
context data in hybrid system structures where infrastructure-based and ad-hoc<br />
networks are integrated. Th e acquisition of data is based on mobile devices such as<br />
mobile phones and their integrated sensors. Th ese devices opportunistically read environmental<br />
conditions and communicate the readings.<br />
SFB 627 – Project B5 – Dynamic, Time-referenced Model Data<br />
Personnel: R. Lange<br />
F<strong>und</strong>ed by: Deutsche Forschungsgemeinschaft (DFG)<br />
Duration: 1.1.2007 – 31.12.2010<br />
Th is project focuses on the management and retrieval of highly dynamic context information<br />
(e.g., moving objects’ trajectories). For this purpose, we researched protocols<br />
37
38<br />
for the real-time transmission of trajectory data from moving objects to remote databases<br />
optimizing communication costs and storage consumption [LWG+09,LDR10].<br />
Moreover, we proposed a scalable tree-based structure for indexing context providers<br />
by means of formal descriptions of their partial world models [LDR09].<br />
SFB 627 – Project E3 – Distributed Situation Recognition<br />
Personnel: F. Dürr, S. Rizou<br />
F<strong>und</strong>ed by: Deutsche Forschungsgemeinschaft (DFG)<br />
Duration: 1.1.2007 – 31.12.2010<br />
Project E3 investigates methods for the realization of a distributed situation recognition<br />
system. As a fi rst step, we introduced a system architecture that enables distributed<br />
context reasoning in large-scale scenarios [RHD+10]. Moreover, we developed an<br />
algorithm for the effi cient distribution of the reasoning operators that minimizes the<br />
overall network load [RD+09]. Currently the research is focused on operator placement<br />
strategies that provide Quality of Service guarantees in terms of timeliness to<br />
the users.<br />
Selected Research Projects<br />
SpoVNet 2.0<br />
A basic requirement for the Future Internet is the seamless integration of heterogeneous<br />
devices and access technologies to support a spontaneous deployment and<br />
adaptive execution of distributed applications. Th e SpoVNet project addresses these<br />
requirements by an architecture that<br />
supports overlay-based innovative<br />
communication services to adapt to<br />
the dynamics of the <strong>und</strong>erlying physical<br />
network.<br />
At the IPVS, an exemplary<br />
two-layered overlay-based event service<br />
was developed which supports<br />
adaptation to QoS restrictions and<br />
provides basic security mechanisms.<br />
EONSON is an event notifi cation<br />
service that provides the semantics of<br />
content-based publish/subscribe systems. Service requirements are ensured without<br />
relying on a dedicated broker infrastructure. In addition, Cordies supports the distributed<br />
detection of complex events by providing an expressive correlation description
language and self-organizing distribution algorithms for the placement of correlation<br />
detection operators.<br />
In SpoVNet 2.0, three German universities collaborate to transfer the developed<br />
concepts into applications of the Future Internet. To this end, a scheme to modularize<br />
SpoVNet services and a framework for module composition will be developed. Furthermore,<br />
the project focuses on connectional extensions to SpoVNet that address upcoming<br />
key technologies for the Future Internet such as the Internet of Th ings and<br />
Network Virtualization.<br />
SFB 627 – Project E3 – Distributed Situation Recognition<br />
Th e goal of Project E3 – a joint project of the departments Image Understanding (BV)<br />
and Distributed Systems (VS) of IPVS – is to build a distributed reasoning system.<br />
Our work includes the design of the architecture of the reasoning system, as well as<br />
algorithms both for context reasoning (BV) and for the effi cient distribution of the<br />
reasoning process (VS).<br />
Th e fi gure below shows the architecture of the distributed reasoning system<br />
[RHD+10]. At runtime, situation detection is initialized by the creation of a logical<br />
plan, which describes the detection of a concrete situation at a certain location or for<br />
a given object by an operator graph. Subsequently the system fi nds a mapping of the<br />
operators of the logical plan to physical hosts according to the optimization goal of the<br />
operator placement. Finally, the operators are executed by the selected physical hosts<br />
using the NexusDS operator execution environment.<br />
For the effi cient distribution of the reasoning operators, we have worked on methods<br />
for optimized operator placement to improve the scalability of the distributed reasoning<br />
system. In particular, we have designed a novel distributed operator placement<br />
algorithm [RD+09] that tries to fi nd node coordinates in an intermediate continuous<br />
space, called latency space, such that the bandwidth-delay product of inter-operator<br />
data streams is minimized to optimize the induced network load.<br />
Our current research focuses on the extension of the afore-mentioned place-<br />
mentalgorithm to<br />
include user-defined<br />
restrictions<br />
in terms of<br />
the communication<br />
latency of<br />
data.<br />
Auxiliary y Services<br />
Situa�on Template<br />
Repository<br />
Context Broker<br />
Resource Model<br />
A���������<br />
Context Aware Applica�on<br />
Layer<br />
Request<br />
Feedback<br />
Response p<br />
���text<br />
Rea���i�� Rea���i��<br />
Logical<br />
Planner<br />
Execu�on Environment<br />
Feedback<br />
Layer<br />
Logical Plan<br />
Adapta�on p<br />
Distributed<br />
Operator<br />
�ni�al<br />
Operator Operator<br />
Placement<br />
Physical<br />
Plan<br />
Distributed<br />
Operator<br />
Placement<br />
Execu�on<br />
Observable<br />
Context<br />
World Model<br />
W�rld<br />
Sensor Data Sta�c Data Historical Data Situa�on Data ��del<br />
Layer<br />
Deployment<br />
off<br />
Physical Plan P<br />
Architecture<br />
of<br />
Distributed<br />
Reasoning<br />
System<br />
39
Th e Nexus World<br />
Model connects<br />
the physical and<br />
the virtual world<br />
40<br />
2.6 Joint Activities<br />
Collaborative Research Center 627 NEXUS — Spatial World Models<br />
for Mobile Context-Aware Applications<br />
www.nexus.uni-stuttgart.de<br />
Coordinator: Prof. K. Rothermel<br />
Vice Coordinator: Prof. B. Mitschang, Prof. P. Kühn<br />
Nexus is an interdisciplinary research center in the fi eld of mobile context-aware systems<br />
and applications at the University of <strong>Stuttgart</strong>, f<strong>und</strong>ed by Deutsche Forschungsgemeinschaft<br />
(DFG) since the year 2003. Th e research program of Nexus integrates<br />
a wide range of disciplines, like Computer Science, Geoinformatics, Industrial Manufacturing,<br />
and Philosophy of Science and Technology. Nexus is headed by Prof. Rothermel<br />
from IPVS and consists of 17 research projects including 6 projects from<br />
IPVS (see Section 2.5 for a detailed description of the individual projects). Research<br />
on globally federated context models constitutes the scientifi c core of Nexus, enabling<br />
our vision of a “World Wide Space” – an open system where partial context models<br />
from diff erent providers are integrated into a global and highly detailed spatial world<br />
model.<br />
Physical World<br />
Spatial World Model<br />
Digitale Information Spaces<br />
An important research focus of the current f<strong>und</strong>ing period is quality of context.<br />
As part of this research eff ort Nexus organized the First International Workshop on<br />
Quality of Context (QuaCon 2009), which aimed to bring together researchers from<br />
various fi elds to discuss the diff erent facets of context quality and to make a consolidated<br />
contribution towards an integrated way of treating them. Th e workshop attracted
about 100 researches from Europe, Asia, and from USA. Th is great success also <strong>und</strong>erlines<br />
the importance of research on context quality to the community.<br />
Other major scientifi c meetings of the Nexus project organized with support of<br />
IPVS members in the year 2009 dealt with the industrial application of context-aware<br />
systems (“Industriekolloquium”) and the management and analysis of context information<br />
(Workshop “Verwaltung, Analyse <strong>und</strong> Bereitstellung kontextbasierter Informationen”).<br />
Both were attended by about 80 persons from academia and industry.<br />
As in previous years, Nexus researches from the IPVS present their work on numerous<br />
conferences, symposia and workshops. Among others, prestigious scientifi c<br />
events such as the International Conference on Mobile Data Management (MDM<br />
2009) and the IEEE International Conference on Pervasive Computing and Communications<br />
(PerCom 2009) can be addressed. Moreover, Prof. Rothermel presented<br />
a keynote on “Large-scale Context Management” at PerCom 2010. Also prototype<br />
systems putting our research results into practice were presented at major events like<br />
PerCom 2009 and 2010.<br />
Moreover, research results of Nexus were presented at several public events at the<br />
university. In June 2009 Andreas Brodt presented the NexusWeb demonstrator at the<br />
Science Day (Tag der Wissenschaft ). NexusWeb is a location-aware web application<br />
which utilizes the built-in GPS receiver of a mobile device to show a map of the user’s<br />
surro<strong>und</strong>ings augmented with context data from the Nexus federation in a single web<br />
page [BC09].<br />
In February 2010, Lars Geiger, Jonas Palauro, and Ralph Lange off ered a workshop<br />
for preparatory school students on tracking moving objects. Th e students showed<br />
great interest in the necessary scientifi c backgro<strong>und</strong>s for novel mobile or locationbased<br />
services. At the same occasion, Andreas Brodt off ered a workshop titled “Databases<br />
and Google maps” in which pupils constructed queries on a spatial database<br />
to drive a geo-mashup web application. Both workshops well achieved their goal to<br />
fascinate pupils of the exciting challenges for a new generation of computer science<br />
students.<br />
41
42<br />
2.7 Infrastructure<br />
<strong>Universität</strong> <strong>Stuttgart</strong> - IPVS<br />
Infrastruktur<br />
<strong>Universität</strong>sstraße 38<br />
70569 <strong>Stuttgart</strong><br />
Germany<br />
Scope<br />
Dipl.-Inf. Michael Matthiesen<br />
� Michael.Matthiesen@ipvs.uni-stuttgart.de<br />
� +49-711-7816414<br />
� +49-711-7816248<br />
Th e group “Infrastructure” is responsible for all tasks referring to the administrative<br />
and technical management of the IPVS. At the time the institute had been fo<strong>und</strong>ed,<br />
the responsible persons decided to join the administrative as well as technical resources<br />
in one department to create a skilled and effi cient support team for the institute.<br />
Th e department acts today as a service provider for the fi ve operational research<br />
groups of the institute. Project coordination, administrative management and information<br />
technology support are the major areas of responsibility.<br />
Project Coordination: Th e main focus comprises the support of the acquisition and<br />
management of national as well as European collaborative research activities in basic<br />
and applied research for all the other departments. Among others the administrative<br />
and fi nancial management of the Collaborative Research Center 627 – Nexus is settled<br />
here since 2003. Since 2008 the administrative coordination of three collaborative<br />
projects f<strong>und</strong>ed by the European Community has been added. Th ese tasks as well as<br />
the management of the department itself, the organization of workshops and conferences<br />
or the publication of reports are being taken up by the project coordinator.<br />
Administrative Management: Th e management joins the resource management<br />
(personnel and fi nances) of the institute’s budget as well all third party collaborative<br />
projects. In particular the daily fi nancial accounting transactions, purchases, personnel<br />
administration, travel arrangements and much more are performed.<br />
Information Technology Support: Th e IT support (Computer Lab) is responsible<br />
for the daily operation and maintenance, as well as planning and purchasing the central<br />
IT resources of the institute. Several h<strong>und</strong>reds of computer systems ranging from laptops<br />
and PCs up to compute servers and clusters are managed by the Computer Lab.
Th e local network and the connection to the faculty-wide net is also being administered.<br />
Additionally all the required basic soft ware services in a heterogeneous OS environment<br />
are provided.<br />
Staff<br />
Staff of the<br />
Department:<br />
(from left )<br />
(front)<br />
H. Kniehl,<br />
S. Palmer,<br />
S. Th ielmann,<br />
M. Rasch,<br />
A. Schaupp<br />
(back)<br />
C. Reissner,<br />
R. Aumüller,<br />
F. Fabian,<br />
B. Schusser,<br />
M. Matthiesen<br />
43
Teaching<br />
Number of Diploma Thesis: 30<br />
Number of Master Thesis: 11<br />
Number of Student Thesis: 13<br />
Number of Student Projects ts: 4<br />
44
3 Teaching<br />
www.ipvs.uni-stuttgart.de/lehre<br />
Th e <strong>Institut</strong>e participates in the following study programs:<br />
Diploma in Computer Science: Computer science generally uses formal (mathematical)<br />
and engineering-based methods to solve problems. Both soft ware and hardware<br />
systems form the focus of the program, both as research objectives and as tools.<br />
Diploma in Software Engineering: Th is course is oriented towards the everyday<br />
tasks of soft ware developers and managers in industry. Students learn how to specify,<br />
design, and implement soft ware systems, and are introduced to soft ware project management<br />
and quality assurance. Large projects and team-work are essential aspects.<br />
Th is course, which is unique in Germany, has been awarded a national prize and been<br />
given an excellent rating by an international evaluation team.<br />
Msc. in Information Technology: Th is is an international master program. Information<br />
Technology stands for enabling technologies like hard- and soft ware design, electronic<br />
and photonic components, computer and communication systems, communication<br />
networks, multimedia, traffi c and process control systems. It further stands for<br />
innovative solutions in areas like mobile/fi xed communication, Internet, automotive<br />
and industrial automation.<br />
Bsc. in Information Systems: Th e ubiquity of information systems and communications<br />
systems in today‘s economy makes it necessary to have a profo<strong>und</strong> <strong>und</strong>erstanding<br />
of the interrelationships between business processes, organisational design<br />
and the use of IT. Teaching methodological expertise and professional competence for<br />
analysis, design and implementation of IT solutions is the primary aim of the program.<br />
Basics in mathematics and statistics are part of the program as well as classes in social<br />
sciences, economics, business administration and computer science.<br />
Msc. in Information Systems: Th e program off ers a high number of special courses<br />
and course clusters in social sciences, economics, business administration and computer<br />
science to deepen the topics introduced in the Bsc. program.<br />
45
Dissertations<br />
and<br />
Postdoctoral<br />
Qualifi cations<br />
46
4 Dissertations and<br />
Postdoctoral Qualifi cations<br />
Dissertations<br />
Buchheim, Thorsten<br />
Kooperative Verhaltensmodellierung <strong>für</strong> adaptive Multiagentensysteme<br />
Supervisors: Prof. Levi, Prof. Roller<br />
Th e focus of the present dissertation is the modelling of cooperative behavior for<br />
groups of autonomous robots in dynamic environments. Th is is done by a formalism<br />
to describe interactions in a multi robot system in form of so called specialized<br />
interaction networks which defi ne the chronological sequence of distributed<br />
tasks in a group of autonomous robots. Furthermore, this work examines the<br />
application of reinforcement learning to optimize the performance of the group<br />
behavior. Th is is achieved on one hand by improving a single agent’s behavior,<br />
on the other hand by improving the assignment of tasks to the individual robot<br />
agents. Th e latter is achieved by a cost based approach which uses invididual cost<br />
estimates of the agents for performing certain tasks to calculate a cost minimal<br />
assignment of tasks to robots. By approximating these cost estimates through a<br />
reinforcement learning approach the group performance is gradually improved.<br />
Dudkowski, Dominique<br />
F<strong>und</strong>amental storage mechanisms for location based services in mobile<br />
ad hoc networks<br />
Supervisors: Prof. Rothermel, Prof. Marrón<br />
Th e proliferation of mobile wireless communication technology has reached a<br />
considerable magnitude. As of 2009, a large fraction of the people in most industrial<br />
and emerging nations is equipped with mobile phones and other types<br />
of portable devices. Supported by trends in miniaturization and price decline of<br />
electronic components, devices become enhanced with localization technology,<br />
which delivers, via the Global Positioning System, the geographic position to the<br />
user. Th e combination of both trends enables location-based services, bringing information<br />
and services to users based on their whereabouts in the physical world,<br />
for instance, in the form of navigation systems, city information systems, and<br />
47
48<br />
friend locators. A growing number of wireless communication technologies, such<br />
as Wireless Local Area Networks, Bluetooth, and ZigBee, enable mobile devices<br />
to communicate in a purely peer-to-peer fashion, thereby forming mobile ad-hoc<br />
networks. Together with localization technology, these communication technologies<br />
make it feasible, in principle, to implement distributed locationbased<br />
services without relying on any support by infrastructure components. However,<br />
the specifi c characteristics of mobile ad-hoc networks, especially the signifi cant<br />
mobility of user devices and the highly dynamic topology of the network, make<br />
the implementation of locationbased services extremely challenging. Current research<br />
does not provide an adequate answer to how such services can be supported.<br />
Effi cient, robust, and scalable f<strong>und</strong>amental mechanisms that allow for generic<br />
and accurate services are lacking.<br />
Th is dissertation presents a solution to the f<strong>und</strong>amental support of locationbased<br />
services in mobile ad-hoc networks. A conceptual framework is outlined<br />
that implements mechanisms on the levels of routing, data storage, location updating,<br />
and query processing to support and demonstrate the feasibility of location-based<br />
services in mobile ad-hoc networks. Th e fi rst contribution is the concept<br />
of location-centric storage and the implementation of robust routing and<br />
data storage mechanisms in accordance with this concept. Th is part of the framework<br />
provides a solution to the problems of data storage that stem from device<br />
mobility and dynamic network topology. Th e second contribution is a comprehensive<br />
set of algorithms for location updating and the processing of spatial queries,<br />
such as nearest neighbor queries. To address more realistic location-based<br />
application scenarios, we consider the inaccuracy of position information of objects<br />
in the physical world in these algorithms. Extensive analytical and numerical<br />
analyses show that the proposed framework of algorithms possesses the necessary<br />
performance characteristics to allow the deployment of location-based services in<br />
purely infrastructureless networks. A corollary from these results is that currently<br />
feasible location-based services in infrastructure-based networks may be extended<br />
to the infrastructureless case, opening up new business opportunities for service<br />
providers.<br />
Handte, Marcus<br />
System support for adaptive pervasive applications<br />
Supervisors: Prof. Rothermel, Prof. Becker<br />
Driven by the ongoing miniaturization of computer technology as well as the proliferation<br />
of wireless communication technology, Pervasive Computing envisions<br />
seamless and distraction-free task support by distributed applications that are ex-
ecuted on computers embedded in everyday objects. As such, this vision is equally<br />
appealing to the computer industry and the user. Induced by various factors such<br />
as invisible integration, user mobility and computer failures, the resulting computer<br />
systems are heterogeneous, highly dynamic and evolving. As a consequence,<br />
applications that are executed in these systems need to adapt continuously to their<br />
ever-changing execution environment. Without further precautions, the need for<br />
adaptation can complicate application development and utilization which hinders<br />
the realization of the basic vision.<br />
As solution to this dilemma, this dissertation describes the design of system<br />
soft ware for Pervasive Computing that simplifi es the development of adaptive applications.<br />
As opposed to shift ing the responsibility for adapting an application<br />
to the user or the application developer, the system soft ware introduces a component-based<br />
application model that can be confi gured and adapted automatically.<br />
To enable automation at the system level, the application developer specifi es<br />
the dependencies on components and resources in an abstract manner using contracts.<br />
Upon application startup, the system uses the contractual descriptions to<br />
compute and execute valid confi gurations. At runtime, it detects changes to the<br />
confi guration that require adaptation and it reconfi gures the application.<br />
To compute valid confi gurations upon application startup, the dissertation<br />
identifi es the requirements for confi guration algorithms. Based on an analysis of<br />
the problem complexity, the dissertation classifi es possible algorithmic solutions<br />
and it presents an integrated approach for confi guration based on a parallel backtracking<br />
algorithm. Besides from scenario specifi c modifi cations, retrofi tting the<br />
backtracking algorithm requires a problem mapping from confi guration to constraint<br />
satisfaction which can be computed on-the-fl y at runtime. Th e resulting<br />
approach for confi guration is then extended to support the optimization of a cost<br />
function that captures the most relevant cost factors during adaptation. Th is enables<br />
the use of the approach for confi guration upon startup and reconfi guration<br />
during runtime adaptation.<br />
As basis for the evaluation of the system soft ware and the algorithm, the dissertation<br />
outlines a prototypical implementation. Th e prototypical implementation<br />
is used for a thorough evaluation of the presented concepts and algorithms<br />
by means of real world measurements and a number of simulations. Th e evaluation<br />
results suggest that the presented system soft ware can indeed simplify the<br />
development of distributed applications that compensate the heterogeneity, dynamics<br />
and evolution of the <strong>und</strong>erlying system. Furthermore, they indicate that<br />
the algorithm for confi guration and the extensions for adaptation provide a suffi<br />
ciently high performance in typical applications scenarios. Moreover, the results<br />
also suggest that they are preferable over of alternative solutions.<br />
49
50<br />
To position the presented solution within the space of possible and existing<br />
solutions, the dissertation discusses major representatives of existing systems and<br />
it proposes a classifi cation of the relevant aspects. Th e relevant aspects are the <strong>und</strong>erlying<br />
conceptual model of the system and the distribution of the responsibility<br />
for confi guration and adaptation. Th e classifi cation <strong>und</strong>erlines that in contrast<br />
to other solutions, the presented solution provides a higher degree of automation<br />
without relying on the availability of a powerful computer. Th us, it simplifi<br />
es the task of the application developer without distracting the user while being<br />
applicable to a broader range of scenarios. Aft er discussing the related approaches<br />
and clarifying similarities and diff erences, the dissertation concludes with a short<br />
summary and an outlook on future work.<br />
Kraft, Tobias<br />
Optimization of query sequences.<br />
Supervisors: Prof. Mitschang, Prof. Härder<br />
Query optimization is a well-known topic in database research since the 1970s.<br />
Th is thesis highlights a special area of query optimization that arises from new<br />
trends in the usage of databases. Whereas in the beginning databases were primarily<br />
used for transaction-oriented processing of operative data, today databases are<br />
also used to facilitate reporting and analysis on consolidated, historic data. For<br />
the latter, the data is loaded into a large data warehouse and aft erwards it is being<br />
analyzed by the use of tools. Th e tools used to model the fl ows that extract the<br />
operative data from the source systems, transform these data and load it into the<br />
data warehouse as well as the tools that process the data stored in the data warehouse<br />
oft en generate sequences of SQL statements that break down a complex<br />
fl ow or request into a sequence of computational steps. Th e optimization of this<br />
kind of sequences with respect to runtime is the focus of this thesis. We propose<br />
a heuristic as well as a cost-based approach for this optimization problem. Th e<br />
cost-based approach is just an enhancement of the heuristic approach. It results<br />
from adding a cost estimation component to the optimizer architecture of the<br />
heuristic approach and by replacing the heuristic control strategy by a control<br />
strategy that considers cost estimates. Both approaches are rule-based approaches<br />
that rewrite a given sequence of SQL statements into a syntactically diff erent but<br />
semantically equivalent sequence of SQL statements. Th erefore, we specify a set<br />
of rewrite rules. For cost estimation, we employ the capabilities of the query optimizer<br />
of the <strong>und</strong>erlying database management system (DBMS) which is responsible<br />
for the execution of the query sequences. To improve the quality of these
cost estimates, we support the query optimizer of the <strong>und</strong>erlying DBMS with statistics<br />
that we derive from histogram propagation. For this purpose, we need an<br />
interface that allows to access and manipulate statistics in the <strong>und</strong>erlying DBMS.<br />
Since there exists no standardized interface for this purpose, we defi ne our own<br />
DBMS-independent interface. For the heuristic approach as well as for the costbased<br />
approach, we provide prototypic implementations in JAVA. Furthermore,<br />
we have implemented the DBMS-independent interface for the three commercial<br />
DBMSs IBM DB2, Oracle, and Microsoft SQL Server. We report on the results<br />
of experiments that we conducted with our prototypes and some sample<br />
sequences that we derived by using a commercial tool for online analytical processing<br />
(OLAP). Th ey show the eff ectiveness of our optimization approach and<br />
they highlight the optimization potential that lies in rewriting sequences of SQL<br />
statements. Finally, we draw a conclusion and suggest some interesting points for<br />
future research.<br />
Lu, Jing<br />
Constraints and triggers to enhance XML-based data integration systems<br />
Supervisors: Prof. Mitschang, Prof. Deßloch<br />
XML is becoming one of the main technological integredients of the Internet. It<br />
is now accepted as the standard for information exchange. XML-based data integration<br />
system, which enables sharing and cooperation with legacy data sources,<br />
arises as a more and more important data service provider on the web. Th ese services<br />
can provide the users with a uniform interface to a multitude of data sources<br />
such as relational databases, XML fi les, text fi les, delimited fi les, Excel fi les, etc.<br />
Users can thus focus on what they want, rather than think about how to obtain<br />
the answers. Th erefore, users do not have to carry on the tedious tasks such as<br />
fi nding the relevant data sources, interacting with each data source in isolation using<br />
the local interface and combining data from multiple data sources.<br />
Users are always expecting better query performance and data consistency<br />
from the data integration systems. Th is work proposes an approach to support<br />
constraints and triggers in the XML-based data integration system in order to<br />
optimize queries and to enforce data consistency. Constraints and triggers have<br />
long been recognized to be useful in semantic query optimization and data consistency<br />
enforcement in relational databases. Th is work fi rst gives an approach to use<br />
constraints from the heterogeneous data sources to semantically optimize queries<br />
submitted to the XML-based data integration system. Diff erent constraints from<br />
51
52<br />
the data sources are fi rst integrated into a uniform constraint model. Th en the<br />
constraints in the uniform constraint model are stored in the constraint repository.<br />
Traditional semantic query optimization techniques in the relational database<br />
are analyzed and three of them are reused and applied by the semantic query<br />
optimizer for XML-based data integration system. Among them are detection of<br />
empty results, join elimination and predicate elimination. Performance is analyzed<br />
according to the data source type and the data volume. Th e semantic query<br />
optimizer works best when the data sources are non-relational, the data volume is<br />
huge and the execution cost is expected to be high.<br />
In order to make the XML-based data integration system fully equipped<br />
with data manipulation capabilities, programming frameworks which support<br />
update at the integration level are being developed. Th is work discusses how to<br />
realize update in the XML-based data integration system <strong>und</strong>er the Service Data<br />
Objects programming framework. When the user is permitted to submit updates,<br />
it is necessary to guarantee data integrity and enforce active business logics in the<br />
data integration system. Th is work presents an approach by which active rules including<br />
integrity constraints are enforced by XQuery triggers. An XQuery trigger<br />
model in conformance to XQuery update model proposed by W3C is defi<br />
ned. How to defi ne active rules and integrity constraints by XQuery triggers<br />
is discussed. Triggers and constraints are stored in the trigger repository. Th e architecture<br />
supporting XQuery trigger service in the XML-based data integration<br />
system is proposed. Important components including event detection, trigger<br />
scheduling, condition evaluation, action fi ring and trigger termination are discussed.<br />
Th e whole XQuery trigger service architecture above a data integration<br />
system is implemented in BEA AquaLogic DataService Platform <strong>und</strong>er the Service<br />
Data Objects programming framework. Experiments show active rules and<br />
integrity constraints are enforced easily, effi ciently and conveniently at the global<br />
level.<br />
Constraints and triggers play an important role in XML-based data integration<br />
systems. Using constraints and triggers in the XML-based data integration<br />
system we can effi ciently improve query performance and enforce data consistency.
Publications<br />
54
5 Publications<br />
Books / Book Chapters<br />
[K10a] Kernbach, S.: Challenges of Pervasive Adaptation: Viewpoint<br />
from Robotics. In “Pervasive Adaptation. Research Agenda for Future<br />
and Emerging Technologies”. A book by Th . Sc. Community<br />
(Coordinated by A. Ferscha), PerAda, 2010, pp. 26-29, doi<br />
[K10b] Kernbach, S.: Concepts of Symbiotic Robot Organisms. In “Symbiotic<br />
Multi-Robot Organisms” (Eds.: P. Levi and S. Kernbach),<br />
pp. 5-24, Springer-Verlag, 2010<br />
[K10c] Kernbach, S.: Towards application of collective robotics in industrial<br />
environment (chapter in book), In G.G. Rigatos (ed.), Industrial<br />
Systems: Modelling, Automation and Adaptive Behaviour,<br />
IGI Global, 2010<br />
[KK10] Kernbach, S.; Kernbach, O.: Structural Self-Organized Control. In<br />
“Symbiotic Multi-Robot Organisms” (Eds.: P. Levi and S. Kernbach),<br />
pp. 309-328, Springer-Verlag, 2010<br />
[M10a] Meister, E.: Kinematics and Dynamics for Robot Organisms. In<br />
“Symbiotic Multi-Robot Organisms” (Eds.: P. Levi and S. Kernbach),<br />
pp. 329-339, Springer-Verlag, 2010<br />
[M10b] Meister, E.: Pervasive Adaptation in Robotic Systems. In “Pervasive<br />
Adaptation. Research Agenda for Future and Emerging Technologies”.<br />
A book by Th . Sc. Community (Coordinated by A. Ferscha),<br />
PerAda, 2010, pp. 10-11<br />
[MZS+09] Monteiro, R. S.; Zimbrão, G.; de Souza, J. M.; Schwarz, H.;<br />
Mitschang, B.: Exploring Calendar-based Pattern Mining in Data<br />
Streams. In: Nguyen, Th o Manh (ed.): Complex Data Warehousing<br />
and Knowledge Discovery for Advanced Retrieval Development:<br />
Innovative Methods and Applications, IGI Global, 2009<br />
55
56<br />
[RK10] Redko, V.; Kernbach, S.: Self-learning Behavior of Virus-like Artifi<br />
cial Organisms. In Symbiotic Multi-Robot Organisms (Eds.: P.<br />
Levi and S. Kernbach), pp. 409-429, Springer-Verlag, 2010<br />
Articles in Journals<br />
[BBS+10] Bader, M.; Böck, C.; Schwaiger, J.; Vigh, C. A.: Dynamically Adaptive<br />
Simulations with Minimal Memory Requirement - Solving the<br />
Shallow Water Equations Using Sierpinski Curves. SIAM Journal<br />
of Scientifi c Computing 32(1), p. 212–228, February 2010<br />
[BKZ+09] Benkmann, R.; Käppeler, U.-P.; Zweigle, O.; Lafrenz, R.; Levi, P.:<br />
Resolving Inconsistencies Using Multi-Agent Sensor Systems. In:<br />
Noberto Pires, J. (ed.): Robotica. Vol. 03/09(76), Coimbra: Engebook,<br />
2009<br />
[BKR09] Briones, J. A.; Koldehofe, B.; Rothermel, K.: Adaptive Publish/<br />
Subscribe for Wireless Mesh Networks. In: SPINE : 2009<br />
[DB09] Dangelmayr, C.; Blochinger, W.: Aspect-oriented component assembly<br />
- a case study in parallel soft ware design. Soft ware: Practice<br />
and Experience, Vol. 39(9), 2009<br />
[GMK+09] Gauger, M.; Marrón, P. J.; Kauker, D.; Rothermel, K.: Low Overhead<br />
Assignment of Symbolic Coordinates in Sensor Networks.<br />
In: Telecommunication Systems, 2009<br />
[HB09] Held, M.; Blochinger, W.: Structured collaborative workfl ow design.<br />
Future Generation Computer Systems - Th e International<br />
Journal of Grid Computing: Th eory, Methods and Applications,<br />
Vol. 25(6), 2009<br />
[KTK+09] Kernbach, S.; Th enius, R.; Kernbach, O.; Schmickl., T.: Re-Embodiment<br />
of Honeybee Aggregation Behavior in Artifi cial Micro-<br />
Robotic System, In “Adaptive Behavior”, Vol. 17(3), pp. 237-259,<br />
2009
[LB09] Lu, J.; Mitschang, B.: Enforcing Data Consistency in Data Integration<br />
Systems by XQuery Trigger Service. In: International Journal<br />
of Web Information Systems. Vol. 5(2), Emerald Group Publishing<br />
Limited, 2009<br />
[SPR09] Schilling, B.; Pletat, U.; Rothermel, K.: Event Correlation in Heterogeneous<br />
Environments. In: it - Information Technology - Complex<br />
Event Processing, 2009<br />
[SBP10] Schulz, S.; Blochinger, W.; Poths, M.: Orbweb - a network substrate<br />
for peer-to-peer grid computing based on open standards.<br />
Journal of Grid Computing, Vol. 8(1), 2010<br />
[SBH09] Schulz, S.; Blochinger, W.; Hannak, H.: Capability-aware information<br />
aggregation in peer-to-peer grids - methods, architecture, and<br />
implementation. Journal of Grid Computing, Vol. 7(2), 2009<br />
Articles in Proceedings<br />
[AKM+09] Alves, S.; Koldehofe, B.; Miranda, H.; Taiani, F.: Design of a Backup<br />
Network for Catastrophe Scenarios. In: Proceedings of the<br />
ACM International Workshop on Advanced Topics in Mobile<br />
Computing for Emergency Management: Communication and<br />
Computing Platforms (MCEM), 2009<br />
[AES09] Avrutin, V.; Eckstein, B.; Schanz, M.: Th e bandcount increment<br />
scenario. III. Deformed structures. In: Proceedings of the Royal<br />
Society A. Vol. 465(2101), Royal Society Publishing, 2009<br />
[BS09] Brodt, A.; Stach, C.: Mobile ortsbasierte Browserspiele. In: Gesellschaft<br />
<strong>für</strong> Informatik e.V. (Hrsg) (ed.): Tagungsband der 39. GI-<br />
Jahrestagung, 28.9. - 2.10.2009, <strong>Universität</strong> zu Lübeck<br />
57
58<br />
[CEB+09] Cipriani, N.; Eissele, M.; Brodt, A.; Großmann, M.; Mitschang, B.:<br />
NexusDS: A Flexible and Extensible Middleware for Distributed<br />
Stream Processing. In: ACM (ed.): IDEAS ‘09: Proceedings of the<br />
2008 International Symposium on Database Engineering & Applications,<br />
2009<br />
[CL09] Cipriani, N.; Lübbe, C.: Ausnutzung von Restriktionen zur Verbesserung<br />
des Deployment-Vorgangs des <strong>Verteilte</strong>n Datenstromverarbeitungssystems<br />
NexusDS. In: Fischer, Stefan (ed.); Maehle,<br />
Erik (ed.); Reischug, Rüdiger (ed.): Beiträge der 39. Jahrestagung<br />
der Gesellschaft <strong>für</strong> Informatik e.V. (GI), 2009<br />
[EFH+09] Eberle, H.; Föll, S.; Herrmann, K.; Leymann, F.; Marconi, A.; Unger,<br />
T.; Wolf, H.: Enforcement from the Inside: Improving Quality<br />
of Business in Process Management. In: 2009 IEEE International<br />
Conference on Web Services (ICWS 2009), 2009<br />
[GDR09] Geiger, L.; Dürr, F.; Rothermel, K.: On Contextcast: A Contextaware<br />
Communication Mechanism. In: IEEE International Conference<br />
on Communications, 2009. ICC ‘09., 2009<br />
[GHL+09] Grossmann, M.; Hönle, N.; Lübbe, C.; Weinschrott, H.: An Abstract<br />
Processing Model for the Quality of Context Data. In: Proceedings<br />
of the 1st International Workshop on Quality of Context,<br />
2009<br />
[GHR09] Grau, A.; Herrmann, K.; Rothermel, K.: Effi cient and Scalable<br />
Network Emulation using Adaptive Virtual Time. In: Proceedings<br />
of the 18th International Conference on Computer Communications<br />
and Networks (ICCCN 2009), 2009<br />
[GSD+09] Geiger, L.; Schertle, R.; Dürr, F.; Rothermel, K.: Temporal Addressing<br />
for Mobile Context-Aware Communication. In: Proceedings<br />
of the Sixth Annual International Conference on Mobile and<br />
Ubiquitous Systems: Computing, Networking and Services (MobiQuitous<br />
‘09), Toronto, ON, Canada, July 13-16, 2009
[HBW09] Held, M.; Blochinger, W.; Werning, M.: Bioinformatics Workfl<br />
ows with Calvin. In 10th International Conference on Web Information<br />
Systems Engineering (WISE), number 5802 in LNCS,<br />
2009, Springer-Verlag<br />
[KBE+09] Koch, A.; Berthelot, A.; Eckstein, B.; Zweigle, O.; Häussermann,<br />
K; Käppeler, U.-P.; Tamke, A.; Rajaie, H.; Levi, P.: Advanced data<br />
logging in RoboCup, In Proceedings of the AMS 2009, Autonome<br />
Mobile <strong>Systeme</strong>, Karlsruhe, Germany, December 3-4, 2009, Informatik<br />
Aktuell: 1-8<br />
[KGS+09a] Käppeler, U.-P.; Gerhardt, A.; Schieberle, C.; Wiselka, M.;<br />
Häussermann, K.; Zweigle, O.; Levi, P.: Reliable Situation Recognition<br />
based on Noise Levels. In Proceedings of the First International<br />
Conference on Disaster Management and Human Health<br />
Risk, New Forest, September 23-25, 2009, volume 110 of WIT<br />
Transactions on the Built Environment, pages 127–137, New Forest,<br />
UK, September 2009<br />
[KGS+09b] Käppeler, U.-P.; Gerhardt, A.; Schieberle, C; Wiselka, M.;<br />
Häussermann, K.; Zweigle, O.; Levi, P.: Reliable situation recognition<br />
based on noise levels. In Proceedings of the First International<br />
Conference on Disaster Management and Human Health Risk,<br />
2009<br />
[KLM+09] Kernbach, S.; Levi, P.; Meister, E.; Schlachter, F.; Kernbach, O.:<br />
Towards self-adaptation of robot organisms with a high developmental<br />
plasticity. In Proc. of the First International Conference<br />
on Adaptive and Self-adaptive Systems and Applications (ADAP-<br />
TIVE-09), Athens/Glyfada, Greece, 2009, pp. 180-187, doi:<br />
10.1109/ComputationWorld.2009.11<br />
[KMS+09] Kernbach, S.; Meister, E.; Scholz, O.; Humza, R.; Liedke, J.; Ricotti,<br />
L.; Jemai, J.; Havlik, J.; Liu, W.: Evolutionary Robotics: Th e<br />
Next-Generation-Platform for On-line and On-board Artifi cial<br />
Evolution. In Proc. of the IEEE Congress on Evolutionary Computation<br />
(IEEE CEC-2009), May 18-21, Trondheim, Norway,<br />
2009, pp. 1079-1086, doi: 10.1109/CEC.2009<br />
59
60<br />
[KHS+09] Kernbach, S; Hamann, H.; Stradner, J.; Th enius, R.; Schmickl, T.;<br />
Crailsheim, K.; van Rossum, A. C.; Sebag, M.; Bredeche, N.; Yao,<br />
Y.; Baele, G.; Van de Peer, Y.; Timmis, J.; Mokhtar, M.; Tyrrell, A.;<br />
Eiben, A. E.; McKibbin, S. P.; Liu, W.; Winfi eld, A. F. T.: On adaptive<br />
self-organization in artifi cial robot organisms. In Proc. of the<br />
2009 Computation World: Future Computing, Service Computation,<br />
Cognitive, Adaptive, Content, Patterns (ComputationWorld<br />
2009), pp. 33-43, November 15-20, Athens, Greece, 2009. doi:<br />
10.1109/ ComputationWorld.2009.9<br />
[KMM+09] Koldehofe, B.; Mogensen, M.; Monod, M.; Quéma, V.: Heterogeneous<br />
Gossip. In: Proceedings of the 10th ACM/IFIP/USENIX<br />
International Conference on Middleware, 2009<br />
[LBC+10] Lübbe, C.; Brodt, A.; Cipriani, N.; Sanft mann, H.: NexusVIS: A<br />
Distributed Visualization Toolkit for Mobile Applications (Demonstration).<br />
In: Proceedings of the 8th Annual IEEE International<br />
Conference on Pervasive Computing and Communications (Per-<br />
Com ‘10); Mannheim, Germany, March 2010<br />
[LDR10] Lange, R.; Dürr, F.; Rothermel, K.: Effi cient Tracking of Moving<br />
Objects using Generic Remote Trajectory Simplifi cation (Demo<br />
Paper). In: Proceedings of the 8th Annual IEEE International<br />
Conference on Pervasive Computing and Communications<br />
- Workshops (PerCom Workshops ‘10). Mannheim, Germany.<br />
March 2010<br />
[LWG+09] Lange, R.; Weinschrott, H.; Geiger, L.; Blessing, A.; Dürr, F.; Rothermel,<br />
K.; Schütze, H.: On a Generic Uncertainty Model for Position<br />
Information. In: Proceedings of the 1st International Workshop<br />
on Quality of Context (QuaCon ’09), <strong>Stuttgart</strong>, Germany,<br />
June 2009<br />
[MJH+09] Minguez, J.; Jakob, M.; Heinkel, U.; Mitschang, B.: A SOA-based<br />
approach for the integration of a data propagation system. In:<br />
IRI’09: Proceedings of the 10th IEEE international conference on<br />
Information Reuse & Integration, 2009
[MTD+09] Memon, F.; Tiebler, D.; Dürr, F.; Rothermel, K.; Tomsu,M.; Domschitz,<br />
P.: Scalable Spatial Information Discovery over Distributed<br />
Hash Tables. In: Procs. of 4th International Conference on<br />
COMmunication System soft WAre and middlewaRE (COM-<br />
SWARE’09), Dublin, Ireland, June 2009. ACM., 2009<br />
[RFB+09] Rothermel, K. (Hrsg); Fritsch, D. (Hrsg); Blochinger, W. (Hrsg);<br />
Dürr, F. (Hrsg): Proceedings of the First International Workshop<br />
on Quality of Context (QuaCon 2009); <strong>Stuttgart</strong>, Germany, June<br />
25-26, 2009, Springer, 2009<br />
[RHD+10] Rizou, S.; Häussermann, K.; Dürr, F.; Cipriani, N.; Rothermel, K.:<br />
A System for Distributed Context Reasoning. In: Proceedings of<br />
the Sixth International Conference on Autonomic and Autonomous<br />
Systems,Cancún, Mexico, March 7-13, 2010<br />
[RM09] Radeschütz, S.; Mitschang, B.: Extended Analysis Techniques For<br />
a Comprehensive Business Process Optimization. In: Proc. of the<br />
International Conference on Knowledge Management and Information<br />
Sharing (KMIS 2009), Portugal, 6.-8. Oktober, 2009<br />
[RNB10] Radeschütz, S.; Niedermann, F.; Bischoff , W.: BIAEditor - Matching<br />
Process and Operational Data for a Business Impact Analysis.<br />
In: Proc. of the 13th International Conference on Extending Database<br />
Technology (EDBT 2010), Switzerland, March 22-26, 2010<br />
[SMK+09] Schlachter, F.; Meister, E.; Kernbach, S.; Levi, P.: Employing<br />
symbiotic robots to enhance pervasive systems. PerAda Magazin<br />
(www.perada-magazine.eu/pdf/1499/1499.pdf ), 2009, doi:<br />
10.2417/2200902.1499<br />
[WDR09] Weinschrott, H.; Dürr, F.; Rothermel, K.: Effi cient Capturing of<br />
Environmental Data with Mobile RFID Readers. In: Proceedings<br />
of the 10th International Conference on Mobile Data Management<br />
(MDM’09), 2009<br />
[WHR09] Wolf, H.; Herrmann, K.; Rothermel, K.: Modeling Dynamic Context<br />
Awareness for Situated Workfl ows. In: OTM 2009 Workshops,<br />
2009<br />
61
62<br />
[WKL+09] Wieland, M.; Käppeler, U.-P.; Levi, P.; Leymann, F.; Nicklas, D.:<br />
Towards Integration of Uncertain Sensor Data into Context-aware<br />
Workfl ows. In GI-Edition Lecture Notes in Informatics (LNI),<br />
editor, Tagungsband INFORMATIK 2009 Im Focus das Leben,<br />
39. Jahrestagung der Gesellschaft <strong>für</strong> Informatik e.V. (GI), Lübeck,<br />
September 2009<br />
[ZHK+09a] Zweigle, O.; Häussermann, K.; Käppeler, U.-P.; Levi, P.; Extended<br />
TA Algorithm for adapting a Situation Ontology. In Proceedings<br />
of the FIRA RoboWorld Congress 2009, Progress in Robotics, volume<br />
44 of Communications in Computer and Information Science,<br />
pages 364-371, Incheon, Korea, August 2009<br />
[ZHK+09b] Zweigle, O.; Häussermann, K.; Käppeler, U.-P.; Levi, P.: Supervised<br />
learning algorithm for automatic adaption of situation templates<br />
using uncertain data. In ICIS ’09: Proceedings of the 2nd International<br />
Conference on Interaction Sciences, pages 197-200, New<br />
York, NY, USA, 2009. ACM<br />
[ZKR+09] Zweigle, O.; Käppeler, U.-P.; Rajaie, H.; Häussermann, K.; Tamke,<br />
A.; Koch, A.; Eckstein, B.; Aichele, F.; Levi, P.: 1. RFC <strong>Stuttgart</strong><br />
Team Description 2009. In: TU Graz (ed.): RoboCup 2009 International<br />
Symposium<br />
[ZMR+09] Zweigle, O.; van de Molengraft , R.; d’Andrea, R.; Häussermann,<br />
K.: RoboEarth: connecting robots worldwide. In ICIS ‘09: Proceedings<br />
of the 2nd International Conference on Interaction<br />
Sciences,pages 184-191, New York, NY, USA, 2009. ACM<br />
Others<br />
[BBB+10] Barisch, M.; Berg, S.; Brodt, A.; Geiger, L.; Gerpott, T.; Gutscher,<br />
A.; Hubig, Ch.; Kühlewind, M.; Kühn, P. J.; Siemoneit, O.: Aspekte<br />
der Abrechenbarkeit <strong>und</strong> Bepreisung kontextbezogener <strong>Systeme</strong>.<br />
Sonderforschungsbereich SFB 627 Bericht 2010/03, 2010
[BBD+10] Becker, S.; Blessing, A.; Dürr, F.; Geiger, L.; Großmann, M.;<br />
Gutscher, A.; Häussermann, K.; Heesen, J.; Käppeler, U.-P.; Lange,<br />
R.; Peter, M.; Siemoneit, O.; Weinschrott, H.; Zweigle, O.; Levi, P.;<br />
Rothermel, K.: Reference Model for the Quality of Context Information.<br />
Sonderforschungsbereich SFB 627 Bericht 2010/02, 2010<br />
[BK09] Buchmann, A. (Hrsg); Koldehofe, B. (Hrsg): IT-Information<br />
Technology, Vol. 51(2009) 5. Complex Event Processing. Oldenbourg<br />
Verlag, Oktober 2009<br />
[GDR10] Geiger, L.; Dürr, F.; Rothermel, K.: Aggregation of User Contexts<br />
in Context-based Communication. Sonderforschungsbereich SFB<br />
627 Bericht 2010/01, 2010<br />
[GBG10] Gerpott, T.; Berg, S.; Geiger, L.: Nutzung standortbezogener Mobilfunkdienste<br />
bei Privatk<strong>und</strong>en in Deutschland. Sonderforschungsbereich<br />
SFB 627 Bericht 2010/04, 2010<br />
[K09] Käppeler, U.-P.: Metriken <strong>für</strong> Qualitätskennzahlen zur Degradierung<br />
von skalaren Sensordaten. <strong>Universität</strong> <strong>Stuttgart</strong>: Sonderforschungsbereich<br />
SFB 627, Technischer Bericht Nr. 2009/02, 16<br />
Seiten, 2009<br />
[LDR09] Lange, R.; Dürr, F.; Rothermel, K.: Indexing source descriptions<br />
for large-scale distributed ontology-based information systems.<br />
Technical Report SFB 627 Bericht Nr. 2009/04, 2009<br />
[RDR+09] Rizou, S.; Dürr, F.; Rothermel, K.: Solving the Multi-operator<br />
Placement Problem in Large-Scale Operator Networks, Technical<br />
Report SFB 627 Bericht Nr. 2009/03, 2009<br />
63
Events<br />
64
6 Events<br />
First International Workshop on Quality of Context<br />
(QuaCon 2009)<br />
http://www.quacon09.nexus.uni-stuttgart.de/<br />
On the 25th and 26th of June 2009, the Collaborative Research Center Nexus organized<br />
the First International Workshop on Quality of Context (QuaCon 2009) at<br />
<strong>Universität</strong> <strong>Stuttgart</strong>. QuaCon 2009 was the fi rst international scientifi c meeting that<br />
specifi cally focused on the diff erent aspects of quality of context data. It aimed to bring<br />
together researchers from various fi elds to discuss diff erent aspects of context quality<br />
and to make a consolidated contribution towards an integrated way of treating this<br />
topic. Th e workshop received 19 high quality paper submissions of researchers from<br />
Europe, USA, and Asia. Th e international program committee selected 11 papers for<br />
presentation. Additionally, 5 invited contributions by the following internationally renowned<br />
experts in the fi eld have been included into the workshop program: Prof. Max<br />
Mühlhäuser (Technical University Darmstadt), Prof. Michael F. Goodchild (University<br />
of California, Santa Barbara), Prof. Johann-Christoph Freytag (HU Berlin), Prof.<br />
Reynold Cheng (University of Hong Kong), and Prof. Christopher Gold (University<br />
of Glamorgan). Th e proceedings of QuaCon 2009 appeared in the Lecture Notes in<br />
Computer Science series [RFBD09].<br />
Nexus Industry Colloquium 2009<br />
(SFB 627)<br />
http://www.nexus.uni-stuttgart.de/i-kolloq09<br />
On 13th of November 2009, the Collaborative Research Center Nexus organized an<br />
industry colloquium at <strong>Universität</strong> <strong>Stuttgart</strong>. At this event high-ranked representatives<br />
of the companies IBM (Dirk Wittkopp), Google (Dr. Mattos) and Th ales ATM<br />
(Dr. Blomenhofer) reported on the practical application of context-aware methods. A<br />
specifi c goal of this colloquium was to evaluate to what extent the results of the f<strong>und</strong>amental<br />
research on context-aware methods conducted by the Nexus project already<br />
fo<strong>und</strong> their way into commercially successful products and applications. Th e Nexus<br />
Industry Colloquium attracted about 90 participants.<br />
65
66<br />
ITPC Day on Cloud Computing<br />
(IBM Technology Partnership Center)<br />
http://itpc.informatik.uni-stuttgart.de<br />
On 3 April 2009, the department of Applications of Parallel and Distributed Systems<br />
was taking part in the ITPC day on “Cloud Computing” that attracted more than 100<br />
participants from university and industry. ITPC stands for IBM Technology Partnership<br />
Center, an industry collaboration between IBM Germany Research & Development<br />
GmbH and the University of <strong>Stuttgart</strong>.<br />
Five speakers reported on current research in cloud computing. Prof. Leymann from<br />
the Univ. <strong>Stuttgart</strong> presented his ideas on “Applications in the Cloud” and Prof.<br />
Mitschang discussed the prominent issues of “Data and Content in the Cloud”. Prof.<br />
Kloeckner, IBM vice president of Cloud Computing Platforms, gave an overview lecture<br />
and introduced the fi eld to the audience. Gerd Breiter, IBM Distinguished Engineer,<br />
presented and discussed architectural settings for cloud computing, and fi nally,<br />
Michael Behrendt, IBM laboratory Böblingen, reported on fi rst practical experiences<br />
in setting up and using cloud technologies in various usage and application scenarios.<br />
IBM Faculty Award for Prof. Mitschang<br />
On 9 October 2009, Prof. Mitschang was awarded the IBM Faculty Award for his<br />
work on “Acceleration Technologies for Interactive Advanced Analytics to Support<br />
New Intelligence for a Smarter Planet”. Th e IBM Faculty Award is a highly recognized<br />
international award for excellent achievements in science and teaching as well as for<br />
building a resilient cooperation between research and industry.<br />
Advertising Computer Science to High School Pupils<br />
Several public events were held at the university to promote Computer Science as an<br />
interesting study program to high school pupils:<br />
▶ In October 2009, Christoph Stach from the department of Applications of Parallel<br />
and Distributed Systems organized a three day workshop for top talented<br />
high school pupils. Th is workshop dealt with pervasive games and how to create<br />
them. Th e upper school students were equipped with Android Smartphones and<br />
learned how to develop their very own location-based game for mobile devices.<br />
▶ In February 2010, Lars Geiger, Jonas Palauro, and Ralph Lange off ered a workshop<br />
for preparatory school students on tracking moving objects. Th e students
showed great interest in the necessary scientifi c backgro<strong>und</strong>s for novel mobile or<br />
location-based services. At the same occasion, Andreas Brodt off ered a workshop<br />
entitled “Databases and Google maps” in which pupils constructed queries on<br />
a spatial database to drive a geo-mashup web application. Both workshops well<br />
achieved their goal to fascinate pupils of the exciting challenges for a new generation<br />
of computer science students.<br />
▶ In February 2010, Prof. Mitschang and PD Holger Schwarz, both from the department<br />
of Applications of Parallel and Distributed Systems, organized a Students<br />
Day for approx. 150 high school students and teachers. Th ere has been a<br />
plenary talk from Google, accompanied by a number of workshops that introduced<br />
the various fi elds on Computer Science to the pupils. Th e goal is to attract<br />
more students to study Computer Science.<br />
RoboCup<br />
(1. Roboter Fußball Club <strong>Stuttgart</strong>)<br />
http://robocup.informatik.uni-stuttgart.de/rfc/www/<br />
For the 1. RFC <strong>Stuttgart</strong>, the cooperative soccer playing robot team of the department<br />
image <strong>und</strong>erstanding, the year 2009 was the most successful in history.<br />
Aft er participating at the RoboCup contests in the middle-size league for 10 years<br />
and several hard-earned 2nd and 3rd places, the team 1. RFC <strong>Stuttgart</strong> was fi nally able<br />
to reach the top and dominated the middle-size leage with a hat-trick of 1st places.<br />
Th e team achieved not only the 1st place of the RoboCup German Open 2009,<br />
held from 20.4.2009 - 24.4.2009 in Hannover, but also the 1st place of the RoboCup<br />
World Championship held from 29.6.2009 - 5.7.2009 in Graz and also there - introducing<br />
a fully autonomous cameraman - the 1st place in the so called RoboCup free<br />
or scientifi c challenge.<br />
During the competition <strong>Stuttgart</strong>’s robots had to make their way against 13 other<br />
teams from eight countries, among them the current world champion Cambada (Portugal).<br />
Besides the teams from Germany, Italy, Th e Netherlands, Portugal, and Austria,<br />
teams from China, Japan, and Iran competed against each other. Aft er the fi nal match<br />
of the competition, the middle-size league robots of the 1. RFC <strong>Stuttgart</strong> - the new<br />
world champion - had to play against the human offi cials of the RoboCup federation.<br />
It turned out, that hereby the robots were the inferior team. Clearly the RoboCup<br />
community has still to bridge a vast distance to reach their fi nal goal to let a humanoid<br />
robot team play against the human world champion by the year 2050.<br />
67
Professional Activities<br />
68
7 Professional Activities<br />
Memberships<br />
Avrutin, Viktor<br />
▶ SICC (Italian Society for Chaos and Complexity)<br />
Hillebrand, Jürgen<br />
▶ EuMA (European Microwave Association)<br />
Levi, Paul<br />
▶ GI (Gesellschaft <strong>für</strong> Informatik)<br />
▶ Faculty Council<br />
▶ Board of Directors IPVS<br />
▶ IFAC<br />
▶ Admission Committee Infotech<br />
▶ Examination Committee Infotech<br />
▶ Director FZI Karlsruhe<br />
▶ Informatik Verb<strong>und</strong> <strong>Stuttgart</strong><br />
Mitschang, Bernhard<br />
▶ Faculty Council<br />
▶ Board of Directors of IPVS<br />
▶ Board of Directors of Graduate School of advanced Manufacturing Engineering<br />
▶ Deputy head of collaborative research center 627<br />
▶ Head of Gemeinsame Kommission Wirtschaft sinformatik<br />
▶ Member of the Technical Committee on Computer Supported Cooperative<br />
Work in Design as part of the Systems, Man, and Cybernetics Society, IEEE<br />
Rothermel, Kurt<br />
▶ Faculty Council<br />
▶ Acting Director of IPVS<br />
▶ Scientifi c Council of IZKT<br />
▶ Coordinator of SFB 627 (Nexus – Spatial World Models for Mobile Context-<br />
Aware Applications)<br />
▶ Editorial Board of the Journal “Informatik: Forschung <strong>und</strong> Entwicklung”<br />
69
70<br />
▶ Editorial Board of the International Journal of Pervasive Computing and Communications<br />
▶ Editorial Board of the Journal “it-information technology”<br />
▶ Principal Investigator in “Graduate School of Advanced Manufacturing Engineering”<br />
(GsAMe), Graduate School of Excellence<br />
▶ Principal Investigator in “Simulation Technology” (SimTech), Center of Excellence<br />
▶ Coordinator of EU FET Project “Adaptive Pervasive Flows” (ALLOW)<br />
▶ Advisory Board of GI Fachgruppe KuVS (Communications and Distributed Systems)<br />
▶ Münchner Kreis<br />
Schanz, Michael<br />
▶ GI (Gesellschaft <strong>für</strong> Informatik),<br />
▶ DPG (Deutsche Physikalische Gesellschaft ),<br />
▶ SICC (Italian Society for Chaos and Complexity)<br />
Simon, Sven<br />
▶ Faculty Council<br />
▶ Board of Directors IPVS<br />
▶ VDE<br />
▶ Informatik Forum <strong>Stuttgart</strong> (infos)<br />
▶ Deutscher Hochschulverband<br />
Program Committees<br />
Mitschang, Bernhard<br />
▶ Tutorial Chair for ACM SIGMOD/PODS Conference, 2010<br />
▶ 8th IEEE International Conference on Pervasive Computing and Communications,<br />
2010<br />
▶ 10th International Conference on Mobile Data Management: Systems, Services<br />
and Middleware, 2009<br />
▶ Conference Co-Chair for Workshop on “Management, Analysis, and Provisioning<br />
of Context-based Information - Verwaltung, Analyse <strong>und</strong> Bereitstellung kontextbasierter<br />
Informationen” as part of 39th Annual GI (Gesellschaft <strong>für</strong> Informatik)<br />
Conference, 2009<br />
▶ 22nd Brazilian Symposium on Database 2009
▶ 12th Int. Conference on Computer Supported Collaborative Work in Design<br />
2009<br />
▶ 13th IFAC Symposium on Information Control Problems in Manufacturing<br />
Rothermel, Kurt<br />
▶ ICDCS 2009, 29th Intern. Conference on Distributed Computing Systems, June<br />
22-26, 2009, Montreal, Quebec, Canada<br />
▶ QuaCon ‚09, 1st Intern. Workshop on Quality of Context, June 09, <strong>Stuttgart</strong>,<br />
Germany<br />
▶ DEBS 2009, 3rd ACM Intern. Conference on Distributed Event-Based Systems,<br />
July 6-9, 2009, Nashville, TN, USA<br />
▶ UIC 2009, Th e Sixth International Conference on Ubiquitous Intelligence and<br />
Computing, July 2009, St Lucia, Brisbane, Australia<br />
▶ INTERNET 2009, Th e First International Conference on Evolving Internet, August<br />
09, Cannes/La Bocca, France<br />
▶ EuroSSC 2009, 4th European Conference on Smart Sensing and Context, September<br />
09, Guildford, UK<br />
▶ IMC 2010, International Conference on Intelligent Interactive Assistance and<br />
Mobile Multimedia Computing 2009, November 2009, Rostock-Warnemünde,<br />
Germany<br />
71
<strong>Universität</strong> <strong>Stuttgart</strong><br />
<strong>Institut</strong> <strong>für</strong> <strong>Parallele</strong> <strong>und</strong><br />
<strong>Verteilte</strong> <strong>Systeme</strong><br />
<strong>Universität</strong>sstraße 38<br />
70569 <strong>Stuttgart</strong><br />
Germany<br />
www.ipvs.uni-stuttgart.de