Completed Projects

This is a list of past projects completed at DOS.

• Stratosphere
  The DFG-funded research unit Stratosphere - Information Management on the Cloud investigated how to run massively data-parallel processing jobs efficiently in Infrastructure-as-a-service clouds.
  Stratosphere is a data parallel, general purpose framework which overcomes the restrictions the map reduce paradigm. It is capable of running batch, as well as real-time tasks like indexing, filtering, transforming, or aggregating data originally found in dataflow systems, but also supports iterations often found in machine learning and graph analysis. It supports structured as well as unstructured data and can run on clusters and in the cloud. To write data analysis programs, it offers a high level, declarative language, a workflow based language on operator basis and a low level language on the graph level.
  With the follow-up project Stratosphere II new features like streaming, iterations and state handling have been researched. Learn more…

• Collaboration with Bundesdruckerei
  Together with the Bundesdruckerei GmbH we looked at establishing trust into IoT data and critical urban infrastructure applications, in particular at new methods for secret-free authentication, relying instead on hardware fingerprints.

• WaterGridSense 4.0
  Water is one of the most important resources to us as human beings and to life in general. As such, there is a great need for managing this resource adequately. Unfortunately, German and international water utility companies, who operate vast water networks, sometimes have surprisingly little knowledge of what is happening within these networks on an ongoing basis. Traditionally, providing such monitoring has been deemed infeasible as the equipment was too expensive to setup and maintain. However, with the introduction of new technologies for sensing, wireless communication, and distributed computing using commodity hardware and open source systems, this goal is finally achievable at a fraction of the cost.

  The BMBF-funded WaterGridSense project was a collaborative effort between water utilities (Hamburg Wasser, Berliner Wasser Betriebe), industry experts (Ingenieurgesellschaft Prof. Dr. Sieker, Walter Tecyard, Funke Kunststoffe, ACO Severin Ahlmann), and research institutions (Technische Universität Berlin, HAW Hamburg). It aimed to design and implement a distributed sensor network in combination with a scalable data analytics platform for providing an online view into the current state of the water networks beneath our cities, as well as, using machine learning techniques to optimize the maintenance of these important facilities.

• OPTIMA
  The demands on urban water networks are constantly growing due to increasing populations, urbanization, and changing climate conditions, where the current climate crisis also leads to more and more extreme weather events. In Berlin, peaks in the load currently require the utilities to relieve about seven billion liters of combined wastewater into the surface waters of the city each year. These combined sewage overflow events lead to an acute surface water pollution and massive fish death. Yet, an expansion of the 150 years old underground sewers would be extremely expensive, if not entirely impossible. Therefore, ways need to be found to maximize the utilization of the existing infrastructure.

  The joint project “OPTIMA” aimed to develop a predictive control system for wastewater pumps that anticipates heavy loads. Data from various sources are used to predict loads in the sewage system. These predictions allow for optimizing the use of the existing retention room of wastewater networks. This way, both damaging backwater and combined sewage overflow events are being reduced.

  The project was funded by the ProFIT program of IBB/ILB and a collaborative effort between partners in Berlin and Brandenburg. Our project partners were the Fluid System Dynamics (FSD) group at TU Berlin, Ingenieurgesellschaft Prof. Dr. Sieker mbH, Fraunhofer FOKUS, and Berliner Wasserbetriebe.

• FLURCA (2020-2022)
  As consumer electronic devices, such as smartphones, become more and more powerful, they need to handle several tasks at the same time as increasingly more services are controlled via them. In the meantime, modern abstraction layers are driving the creation of hiding this complexity from the user while adding technical complexity under the hood. This aggravates the development and maintenance of consumer electronic devices, therefore, poses new challenges for manufacturers and developers. Failures or other unpredictable events can cause severe reputation damage to the manufacturer, especially in cases of unsatisfied consumers. In principle, a failure is created by a defect in the system, which is created by a root cause at runtime in the system and then propagates until it becomes an observable failure. Tracing back the chain from the failure to the root cause is a hard task. The aim of this project was to automatically localize root causes for failures to support developers in an automated manner.

  The project was funded by and in close collaboration with Huawei.

• ide3a (2020-2023)
  The International Alliance for Digital E-learning, E-mobility and E-research in Academia (ide3a) was a collaborative project in the higher education sector involving a multidisciplinary and international consortium. A network of European universities was aiming to develop new tools for the digitalisation of teaching and learning in the internationalisation process, with a particular focus on interdisciplinary research-based learning and short-term mobility in international blended learning formats. We conducted teaching and research on the digitalization of critical infrastructures like water networks, energy grids, sensor networks, and other interconnected urban systems.

  The project was funded by the German Academic Exchange Service (DAAD) supported by the Federal Ministry of Education and Research (BMBF). Our partner universities were the Cracow University of Technology, Dublin City University, Norwegian University of Science and Technology, Politecnico di Milano, and Hasso Plattner Institute.

• Helmholtz-Einstein International Berlin Research School in Data Science (HEIBRiDS)
  The Einstein Center Digital Future (ECDF) and the Helmholtz Association launched HEIBRiDS as a joint graduate program in Data Science. Established in 2018, HEIBRiDS is an interdisciplinary program that trains young scientists in Data Science applications within a broad range of natural science domains, spanning from Earth & Environment, Astronomy, Space & Planetary Research to Geosciences, Materials & Energy and Molecular Medicine. Our project was a collaboration between the TU Berlin and the GFZ Potsdam. In the project, we calibrated the platform magnetometers of non-dedicated satellites flying low Earth orbits, post-launch.

• SYNERGY (2021-2024)
  The SYNERGY project investigated synergies between distributed artificial intelligence and renewable energy generation. The project was a collaboration of DOS and Huawei, funded by the BMBF through Software Campus, resulted in new methods for sustainable federated learning (FedZero, e-Energy’24) and a testbed carbon-aware applications and systems (Vessim, HotCarbon’24).