Auflistung nach:
Auflistung Softwaretechnik-Trends 37(3) - 2017 nach Erscheinungsdatum
1 - 10 von 16
Treffer pro Seite
Sortieroptionen
- ZeitschriftenartikelLeveraging State to Facilitate Separation of Concerns in Reuse-oriented Performance Models(Softwaretechnik-Trends Band 37, Heft 3, 2017) Werle, Dominik; Seifermann, Stephan; Krach, Sebastian D.Each of the five dedicated roles of the Palladio process considers one or more concerns that form a performance prediction model, altogether. Modeling systems that vary their behavior based on a request history, however, requires to break role separation and create dependencies between concerns, thus reducing the reusability of components. Model elements that allow expressing such behavior while maintaining role separation do not exist. We propose a model extension that allows expressing behavior statefully and a transformation to a basic stateless Palladio model. This allows to maintain the role separation and thereby the reusability of components without the need for changes of existing analyses.
- ZeitschriftenartikelNeuer Arbeitskreis Microservices und DevOps in der Fachgruppe Architekturen(Softwaretechnik-Trends Band 37, Heft 3, 2017) Hasselbring, WilhelmAm 31. Mai 2017 fand in Hamburg das Grundungstreffen des neuen Arbeitskreises ”Microservices und DevOps“ in der Fachgruppe Architekturen mit gut zwanzig Teilnehmern statt. Gastgeber war die adesso AG, Niederlassung Hamburg.
- Zeitschriftenartikel8th Symposium on Software Performance (SSP) - Karlsruhe, November 09–10, 2017(Softwaretechnik-Trends Band 37, Heft 3, 2017) Reussner, Ralf; Hasselbring, Wilhelm; Becker, Steffen
- ZeitschriftenartikelTowards Predicting Performance of GPU-dependent Applications on the Example of Machine Learning in Enterprise Applications(Softwaretechnik-Trends Band 37, Heft 3, 2017) Willnecker, Felix; Krcmar, HelmutAlgorithms processed by Graphics Processing Units (GPU) became popular recently. Bitcoin mining algorithms, image processing and all types of machine learning are famous examples for that. Infrastructureas-a-Service provider picked up this trend and offer graphics processing power as part of their service portfolio. The performance gains when choosing a GPU implementation can be enormous. Designing and implementing a GPU-depended algorithm has some fundamental differences compared to classical algorithms, but not all algorithmic problems benefit from GPU usage regarding the overall performance and response time. Especially the interaction between Central Processing Unit (CPU) and GPU must be considered as it can become a bottleneck. Predicting and comparing the performance of GPU-depended applications in combination with their corresponding CPUs allows to assist design decisions in modern applications. In this work, we present concepts on how to predict algorithm performance relying on GPU processing and their relationship with the CPU using the Palladio Component Model and the Palladio Bench.
- ZeitschriftenartikelIs the PCM Ready for ACTORs and Multicore CPUs? — A Use Case-based Evaluation(Softwaretechnik-Trends Band 37, Heft 3, 2017) Frank, Markus; Staude, Stefan; Hilbrich, MarcusMulticore CPUs have been common for years. However, developing parallel software is still an issue. To ease the development, software developers can use a range of frameworks and approaches, e.g., OpenMP, MPI or ACTOR. These approaches have an enormous impact on the performance of the software. Thus, Software Performance Engineering (SPE) needs to consider the impact of the parallelization approaches to deliver reliable results. In this paper, we evaluate the capability of the Palladio Component Model1 (PCM) based on the use case of a bank transaction example with a realization following the ACTOR approach. We observed that the accuracy of the performance predictions is unsatisfying, the modeling is challenging, and the characteristics of the ACTOR approach cannot be modeled. In future we need to consider additional attributes or properties to enrich the PCM as well to include concepts like active resources, message passing, and automatization concepts.
- ZeitschriftenartikelRadarGun: Toward a Performance Testing Framework(Softwaretechnik-Trends Band 37, Heft 3, 2017) Henning, Sören; Wulf, Christian; Hasselbring, WilhelmWe present requirements on a performance testing framework to distinguish it from a functional testing framework and a benchmarking framework. Based on these requirements, we propose such a performance testing framework for Java, called RadarGun. RadarGun can be included into a continuous integration server, such as Jenkins, so that performance tests are executed automatically during the build process. We conducted a feasibility evaluation of this approach by applying it to the continuous integration infrastructure of the Pipe-and-Filter framework TeeTime.
- ZeitschriftenartikelProviding Model-Extraction-as-a-Service for Architectural Performance Models(Softwaretechnik-Trends Band 37, Heft 3, 2017) Walter, Jürgen; Eismann, Simon; Reed, Nikolai; Kounev,SamuelArchitectural performance models can be leveraged to explore performance properties of software systems during design-time and run-time. We see a reluctance from industry to adopt model-based analysis approaches due to the required expertise and modeling effort. Building models from scratch in an editor does not scale for medium and large scale systems in an industrial context. Existing open-source performance model extraction approaches imply significant initial efforts which might be challenging for layman users. To simplify usage, we provide the extraction of architectural performance models based on application monitoring traces as a web service. Model-Extraction-as-a-Service (MEaaS) solves the usability problem and lowers the initial effort of applying model-based analysis approaches.
- ZeitschriftenartikelThe Raspberry Pi: A Platform for Replicable Performance Benchmarks?(Softwaretechnik-Trends Band 37, Heft 3, 2017) Knoche, Holger; Eichelberger, HolgerReplicating results of performance benchmarks can be difficult. A common problem is that researchers often do not have access to identical hardware and software setups. Modern single-board computers like the Raspberry Pi are standardized, cheap, and powerful enough to run many benchmarks, although probably not at the same performance level as desktop or server hardware. In this paper, we use the MooBench micro-benchmark to investigate to what extent Raspberry Pi is suited as a platform for replicable performance benchmarks. We report on our approach to set up and run the experiments as well as the experience that we made.
- ZeitschriftenartikelConverting Traces of In-Memory Database Systems to OPEN.XTRACE on the Example of SAP HANA(Softwaretechnik-Trends Band 37, Heft 3, 2017) Barnert, Maximilian; Streitz, Adrian; Kienegger,Harald; Krcmar, HelmutThe shift of data-intensive application logic to inmemory Database Management Systems increases their importance for the overall performance of the software system. The performance of a processed query on a Database Management System is influenced by the utilized query execution plan, while traces capture the runtime behavior of the processed execution plan. However, the use of proprietary trace formats limits the usability within Application Performance Management tools and Software Performance Engineering approaches. OPEN.XTRACE is an open format to exchange execution traces, but its current data model does not support the integration of internal Database Management System operations. In this paper, we propose a modification to OPEN.XTRACE that enables a common representation of a query execution trace. In addition, we convert traces of the state-of-the-art in-memory Database Management System SAP HANA into this format.
- ZeitschriftenartikelTowards Extracting Realistic User Behavior Models(Softwaretechnik-Trends Band 37, Heft 3, 2017) Jung, Reiner; Adolf, Marc; Dornieden, ChristophWorkloads can be characterized by intensity and user behavior. Combining multiple intensities and behaviors can be used to create workload profiles to evaluate software design and support the prediction of system utilization. The central challenge for workload profiles is their fit to real workloads and in particular the match to specific behaviors. This is especially relevant for understanding and identifying specific user groups and support workload composition by operators. In this paper, we address the identification of such realistic user behaviors utilizing domain specific attributes, evaluate the fitness of potential behavior clustering approaches, and discuss our setup to evaluate further clustering approaches.