A5074535002- Parallel Computing and Optimization Techniques
- Distributed systems and fault tolerance
- Distributed and Parallel Computing Systems
- Software System Performance and Reliability
- Cloud Computing and Resource Management
- Real-Time Systems Scheduling
- Advanced Data Storage Technologies
- Embedded Systems Design Techniques
- Advanced Software Engineering Methodologies
- Service-Oriented Architecture and Web Services
- Software Reliability and Analysis Research
- Algorithms and Data Compression
- Software-Defined Networks and 5G
- Privacy-Preserving Technologies in Data
- Software Testing and Debugging Techniques
- Scientific Computing and Data Management
- IoT and Edge/Fog Computing
- Advanced Malware Detection Techniques
- Cryptography and Data Security
- Cloud Data Security Solutions
- Interconnection Networks and Systems
- Security and Verification in Computing
- Context-Aware Activity Recognition Systems
- Data Stream Mining Techniques
- Radiation Effects in Electronics
University of Potsdam
2015-2024
Hasso Plattner Institute
2015-2024
Humboldt-Universität zu Berlin
1995-2003
Freie Universität Berlin
1995-2002
Software Engineering Institute
2002
In contrast to applications relying on specialized and expensive highly-available infrastructure, the basic approach of microservice architectures achieve fault tolerance - finally high availability is modularize software system into small, self-contained services that are connected via implementation-independent interfaces. Microservices all dependencies deployed environments called containers executed as multiple redundant instances. If a service fails, other instances will often still...
Increasing capabilities of modern microcontrollers greatly increase their applicability to more and complex scenarios. However, unstable ever-changing environmental settings require embedded systems permanently adapt new situations. Dynamic reconfiguration provides a powerful mechanism execute such adaptation strategies. The implementation dynamic is still challenging for real-time control software. Within earlier work we have presented our framework Adapt.NET runtime adaption...
Next generation processor and memory technologies will provide tremendously increasing computing capacities for application scaling. However, this comes at a price: Due to the growing number of transistors shrinking structural sizes, overall system reliability future server systems is about suffer significantly. This makes reactive fault tolerance schemes less appropriate applications under timeliness constraints. We propose an architectural blueprint managing dependability in pro-active...
Microsoft has introduced the .NET architecture as a new component-based programming environment, which allows for easy integration of classical distributed techniques with Web computing. NET defines type system and introduces notions such component, object, interface, are building blocks multi-language applications. As many other component frameworks, mainly focuses on functional interfaces components. Non-functional aspects components, resource usage (CPU, memory), timing behavior,...
Dynamic reconfiguration provides of powerful mechanism to adapt component-based distributed applications changing environmental conditions. We have designed and implemented a framework for dynamic component on the basis Microsoft .NET environment. Within this paper we present an experimental evaluation our infrastructure applications. Our supports description application configurations profiles allows selection particular configuration object/component instantiation based measured In...
Data transfers impose a major bottleneck in heterogenous system architectures. As mitigation strategy, compute resources can be introduced places where data occurs naturally. The increased diversity of turn affects programming models and practicalities software development for near-data kernels raises the question how those made accessible to users applications.
This paper introduces ReDAC, a new algorithm for dynamic reconfiguration of multi-threaded applications. In order to achieve high reliability and availability, distributed component software has support reconfiguration. Typical examples include the application hot-fixes deal with security vulnerabilities. ReDAC can be implemented on top modern component-platforms Java .NET. We extend statical term component, denoting unit deployment, runtime by defining capsule (runtime instance) set...
Todays commercial off-the-shelf computer systems are multicore computing as a combination of CPU, graphic processor (GPU) and custom devices. In comparison with CPU cores, cards capable to execute hundreds up thousands compute units in parallel. To benefit from these GPU resources, applications have be parallelized adapted the target architecture. this paper we show our experience applying NQueens puzzle solution on GPUs using Nvidia's CUDA (Compute Unified Device Architecture) technology....
With the growth of virtualization and cloud computing, more forensic investigations rely on being able to perform live forensics a virtual machine using introspection (VMI). Inspecting through its hypervisor enables investigation without risking contamination evidence, crashing computer, etc. To further access these techniques for investigator/researcher we have developed new VMI monitoring language. This language is based review most commonly used VMI-techniques date, it user monitor...
Using graphics hardware to enhance CPU-based standard desktop applications is a question not only of programming models but also critical optimizations that are required achieve true performance improvements. Two major trends make parallel crucial issue for all software engineers today: the rise many-core CPU architectures and inclusion powerful processing units (CPUs) in every computer. Modern CPUs use ever-increasing transistor count predicted by Moore's law larger caches improved...
Embedded systems have been used to control physical environments for decades. Usually, such use cases require low latencies between commands and actions as well a high predictability of the expected worst-case delay. To achieve this on small, low-powered microcontrollers, Real-Time Operating Systems (RTOSs) are manage different tasks these machines deterministically possible. However, with advent Internet Things (IoT) in industrial applications, same embedded now equipped networking...
Federated learning has the potential to make machine applicable highly privacy-sensitive domains and distributed datasets. In some scenarios, however, a central server for aggregating partial results is not available. fully decentralized learning, network of peer-to-peer nodes collaborates form consensus on global model without trusted party. Often, consists Internet Things (IoT) Edge computing nodes.Previous approaches map gradient batching averaging algorithm from traditional federated...
The carbon footprint of software activities is determined by embodied and operational emissions hardware resources. This paper presents cMemento, a concept that enables operating systems to make carbon-aware memory placement decisions. Main has become heterogeneous in today's computer systems. In addition traditional (and volatile) main (e.g. DRAM), novel technologies with persistent properties are often also available PRAM, FRAM, MRAM). Complementary, there large number new interfaces...
Distributed dataflow systems like MapReduce, Spark, and Flink help users in analyzing large datasets with a set of cluster resources. Performance modeling runtime prediction is then used for automatically allocating resources specific performance goals. However, the actual distributed jobs can vary significantly due to factors interference co-located workloads, varying degrees data locality, failures.We address this problem Ellis, system that allocates an initial target, yet also...
GPU compute devices have become very popular for general purpose computations. However, the SIMD-like hardware of graphics processors is currently not well suited irregular workloads, like searching unbalanced trees. In order to mitigate this drawback, NVIDIA introduced an extension programming models called dynamic parallelism. This enables programs spawn new units work directly on GPU, allowing refinement subsequent items based intermediate results without any involvement main CPU....
Density of medical clinics, ease access to doctors, the age pyramid population - these are differentiating factors among patients living in urban or rural areas which may affect probability survive certain diseases and attacks dramatically. The German Fontane project, a collaboration experts, IT researchers companies, focuses on remote monitoring aftercare facilities for stroke with heart Brandenburg. Within this paper, we discuss central requirements object-oriented self-adaptive middleware...
SUMMARY Over many decades, advances in computer system design and processor manufacturing have resulted an ever‐increasing per‐chip transistor count, which, combination with increased clock frequencies, has led to a tremendous increase single‐thread performance desktop server CPUs. The trend higher integration will continue for the next couple of years. However, count leads additional compute cores within CPU, rather than performance. Programming utilizing these future CPUs impose number...
Federated learning allows a group of distributed clients to train common machine model on private data. The exchange updates is managed either by central entity or in decentralized way, e.g. blockchain. However, the strong generalization across all makes these approaches unsuited for non-independent and identically (non-IID)
Clusters of networked commercial, off the shelf (COTS) workstations are presently used for computation intensive tasks that were typically assigned to parallel computers in past. However, it is hardly possible predict timing behavior such systems or give guarantees about execution times. We show how our SONiC (Shared Objects Net-interconnected Computer) system can control and partitioning a workstation as step towards distributed real time built from COTS components. provides class based...