Cargando…

MASADA: A Modeling and Simulation Automated Data Analysis framework for continuous data-intensive validation of simulation models

Complex networked computer systems are usually subjected to upgrades and enhancements on a continuous basis. Modeling and simulation of such systems helps with guiding their engineering processes, in particular when testing candi- date design alternatives directly on the real system is not an option...

Descripción completa

Detalles Bibliográficos
Autores principales: Foguelman, Daniel Jacob, Bonaventura, Matias Alejandro, Castro, Rodrigo Daniel
Lenguaje:eng
Publicado: 2016
Materias:
Acceso en línea:http://cds.cern.ch/record/2218044
_version_ 1780952136851390464
author Foguelman, Daniel Jacob
Bonaventura, Matias Alejandro
Castro, Rodrigo Daniel
author_facet Foguelman, Daniel Jacob
Bonaventura, Matias Alejandro
Castro, Rodrigo Daniel
author_sort Foguelman, Daniel Jacob
collection CERN
description Complex networked computer systems are usually subjected to upgrades and enhancements on a continuous basis. Modeling and simulation of such systems helps with guiding their engineering processes, in particular when testing candi- date design alternatives directly on the real system is not an option. Models are built and simulation exercises are run guided by specific research and/or design questions. A vast amount of operational conditions for the real system need to be assumed in order to focus on the relevant questions at hand. A typical boundary condition for computer systems is the exogenously imposed workload. Meanwhile, in typical projects huge amounts of monitoring information are logged and stored with the purpose of studying the system’s performance in search for improvements. Also research questions change as systems’ operational conditions vary throughout its lifetime. This context poses many challenges to determine the validity of simulation models. As the behavioral empirical base of the system grows, the question can be posed of whether a model that was deemed valid at one point in time can be regarded as invalid in the context of novel, unprecedented operation conditions. In this work we present a framework and a prototype tool that can help with answering this question in a systematic and automated way. The idea is to select a systems’ operation time interval and get a simulation automatically parameterized, run, and validated, producing a set of results to determine the degree of approximation against available metrics of the real system. We tested our framework and tool in a real case study for TDAQ, the trigger and data acquisition system of the ATLAS particle physics experiment at CERN’s Large Hadron Collider. The results show that with a modest amount of manual set up efforts MASADA can provide a sound mechanism for validating our simulation model for the TDAQ network on a continuous basis, as new evidence becomes available when new collisions take place at CERN’s collider.
id cern-2218044
institution Organización Europea para la Investigación Nuclear
language eng
publishDate 2016
record_format invenio
spelling cern-22180442019-09-30T06:29:59Zhttp://cds.cern.ch/record/2218044engFoguelman, Daniel JacobBonaventura, Matias AlejandroCastro, Rodrigo DanielMASADA: A Modeling and Simulation Automated Data Analysis framework for continuous data-intensive validation of simulation modelsParticle Physics - ExperimentComplex networked computer systems are usually subjected to upgrades and enhancements on a continuous basis. Modeling and simulation of such systems helps with guiding their engineering processes, in particular when testing candi- date design alternatives directly on the real system is not an option. Models are built and simulation exercises are run guided by specific research and/or design questions. A vast amount of operational conditions for the real system need to be assumed in order to focus on the relevant questions at hand. A typical boundary condition for computer systems is the exogenously imposed workload. Meanwhile, in typical projects huge amounts of monitoring information are logged and stored with the purpose of studying the system’s performance in search for improvements. Also research questions change as systems’ operational conditions vary throughout its lifetime. This context poses many challenges to determine the validity of simulation models. As the behavioral empirical base of the system grows, the question can be posed of whether a model that was deemed valid at one point in time can be regarded as invalid in the context of novel, unprecedented operation conditions. In this work we present a framework and a prototype tool that can help with answering this question in a systematic and automated way. The idea is to select a systems’ operation time interval and get a simulation automatically parameterized, run, and validated, producing a set of results to determine the degree of approximation against available metrics of the real system. We tested our framework and tool in a real case study for TDAQ, the trigger and data acquisition system of the ATLAS particle physics experiment at CERN’s Large Hadron Collider. The results show that with a modest amount of manual set up efforts MASADA can provide a sound mechanism for validating our simulation model for the TDAQ network on a continuous basis, as new evidence becomes available when new collisions take place at CERN’s collider.ATL-DAQ-PROC-2016-014oai:cds.cern.ch:22180442016-09-24
spellingShingle Particle Physics - Experiment
Foguelman, Daniel Jacob
Bonaventura, Matias Alejandro
Castro, Rodrigo Daniel
MASADA: A Modeling and Simulation Automated Data Analysis framework for continuous data-intensive validation of simulation models
title MASADA: A Modeling and Simulation Automated Data Analysis framework for continuous data-intensive validation of simulation models
title_full MASADA: A Modeling and Simulation Automated Data Analysis framework for continuous data-intensive validation of simulation models
title_fullStr MASADA: A Modeling and Simulation Automated Data Analysis framework for continuous data-intensive validation of simulation models
title_full_unstemmed MASADA: A Modeling and Simulation Automated Data Analysis framework for continuous data-intensive validation of simulation models
title_short MASADA: A Modeling and Simulation Automated Data Analysis framework for continuous data-intensive validation of simulation models
title_sort masada: a modeling and simulation automated data analysis framework for continuous data-intensive validation of simulation models
topic Particle Physics - Experiment
url http://cds.cern.ch/record/2218044
work_keys_str_mv AT foguelmandanieljacob masadaamodelingandsimulationautomateddataanalysisframeworkforcontinuousdataintensivevalidationofsimulationmodels
AT bonaventuramatiasalejandro masadaamodelingandsimulationautomateddataanalysisframeworkforcontinuousdataintensivevalidationofsimulationmodels
AT castrorodrigodaniel masadaamodelingandsimulationautomateddataanalysisframeworkforcontinuousdataintensivevalidationofsimulationmodels