Logo des Repositoriums

Auflistung nach Schlagwort "modeling techniques"

1 - 3 von 3
  • Zeitschriftenartikel
    Full-Stack Performance Model Evaluation using Probabilistic Garbage Collection Simulation
    (Softwaretechnik-Trends Band 35, Heft 3, 2015) Willnecker, Felix; Koch-Kemper, Bernhard
    Performance models can represent the performance relevant aspects of an enterprise application. Corresponding simulation engines use such models for simulating performance metrics (e.g., response times, resource utilization, throughput) and allow for performance evaluations without load testing the actual system. Creating such models manually often outweighs their benefits. Therefore, recent research created performance model generators, which can generate such models out of Application Performance Management software. However, a full-stack evaluation containing all relevant resources of an enterprise application (Central Processing Unit, memory, network and Hard Disk Drive) has not been conducted to the best of our knowledge. This work closes this gap using a pre-release version of the next generation industry benchmark SPECjEnterpriseNEXT of the Standard Performance Evaluation Corporation as example enterprise application, the Palladio Component Model as performance model and the performance model generator of the RETIT Capacity Manager. Furthermore, this work extends the generated model with a probabilistic garbage collection model to simulate memory allocation and releases more accurately.
  • Zeitschriftenartikel
    Modeling Complex User Behavior with the Palladio Component Model
    (Softwaretechnik-Trends Band 35, Heft 3, 2015) Vögele, Christian; Heinrich, Robert; Heilein, Robert; van Hoorn, André; Krcmar, Helmut
    The specification of workloads is required in order to evaluate performance characteristics of application systems using performance prediction approaches like the Palladio Component Model (PCM). One of the biggest challenges in workload modeling is to ensure that the modeled user behavior adequately resembles the real user behavior. However, PCM offers limited support to model such complex user behavior. Furthermore, reusing modeled activities is not possible. To overcome these limitations, workarounds are required. In order to avoid these workarounds, we extend the meta-model of the PCM Usage Model. We evaluate the extended PCM Usage Model by integrating it into our previous work on automatic extraction of workload specifications. Based on HTTP web logs, recorded from the standard industry benchmark SPECjEnterprise2010, instances of a domain-specific language (DSL) for modeling workload specifications are extracted. Afterwards, these instances are transformed to the extended PCM Usage Model. The evaluation shows that workload characteristics of the simulated workload match the measured workload with high accuracy.
  • Zeitschriftenartikel
    A Performance Model Management Repository Based on the Palladio Component Model
    (Softwaretechnik-Trends Band 35, Heft 3, 2015) Danciu, Alexandru; Brunnert, Andreas; Krcmar, Helmut
    Applying performance models to evaluate component-based enterprise applications in practice is becoming more and more difficult with an increasing organizational complexity. Components can be governed by different organizational units, are subject to a continuous shift between life cycle phases, and exhibit diverging release levels. Creating and maintaining performance models based on the composition of these components can, therefore, represent an elaborate task. A Performance Model Management Repository (PMMR) supports managing performance models in complex enterprise environments. This paper presents the implementation of a PMMR based on the Palladio Component Model (PCM). The PCM meta-model is extended to enable managing and maintaining multiple versions of components and their interfaces. Furthermore, resource demand specifications derived from different hardware environments are integrated into the meta-model. The Palladio-Bench is extended for persisting PMMR elements to EMFStore.