The need for dependable and sustainable products and services hasbeen a topic of major interest in the last decades. Not only should a system be designedto be reliable throughout its lifetime, but it should also be designed to be sustainable,i.e., to minimize the material and energy use, both in the design/constructionstage and during its operation. Stochastic techniques have been widely used to modelsystems characterized by non-deterministic behavior, such as: queuing systems, faulttolerant systems, etc. This kind of models can be used in reliability/availability evaluationunder different operation profiles. In this paper, after a short review and comparisonof different modeling approaches in the field of stochastic modeling, twocase studies are presented. The two cases studies were chosen because of the distributedarchitecture and complexity of the systems: a KNXnet/IP network and apower-Telco network. Through the analysis of these case studies two main aspectsof performability engineering are investigated: the comparison between different operatingstrategies of the KNXnet/IP system and the availability of the service offeredby the Telco nodes. While the first case study offers the possibility of comparingthe efficiency of different package management systems, the second focuses on the definition of an integrated online risk estimator for the reliability of power supply tothe Telco nodes, supported by backup battery systems. We use a modeling languagebased on Stochastic Activity Networks (SAN), a popular formalism that has been inused for modeling a range of complex distributed systems, and Adaptive TransitionSystem, a class of SAN model effective in representing multi component systemsas a set of parallel transition systems. Both the formalisms offer compositionalityprinciples based on building blocks that allow to tackling effectively the complexrelationships of large multi-component systems.
QUANTITATIVE ASSESSMENT OF DISTRIBUTED NETWORKS THROUGH HYBRID STOCHASTIC MODELING
CAVALIERI, Salvatore;Chiacchio F
;
2015-01-01
Abstract
The need for dependable and sustainable products and services hasbeen a topic of major interest in the last decades. Not only should a system be designedto be reliable throughout its lifetime, but it should also be designed to be sustainable,i.e., to minimize the material and energy use, both in the design/constructionstage and during its operation. Stochastic techniques have been widely used to modelsystems characterized by non-deterministic behavior, such as: queuing systems, faulttolerant systems, etc. This kind of models can be used in reliability/availability evaluationunder different operation profiles. In this paper, after a short review and comparisonof different modeling approaches in the field of stochastic modeling, twocase studies are presented. The two cases studies were chosen because of the distributedarchitecture and complexity of the systems: a KNXnet/IP network and apower-Telco network. Through the analysis of these case studies two main aspectsof performability engineering are investigated: the comparison between different operatingstrategies of the KNXnet/IP system and the availability of the service offeredby the Telco nodes. While the first case study offers the possibility of comparingthe efficiency of different package management systems, the second focuses on the definition of an integrated online risk estimator for the reliability of power supply tothe Telco nodes, supported by backup battery systems. We use a modeling languagebased on Stochastic Activity Networks (SAN), a popular formalism that has been inused for modeling a range of complex distributed systems, and Adaptive TransitionSystem, a class of SAN model effective in representing multi component systemsas a set of parallel transition systems. Both the formalisms offer compositionalityprinciples based on building blocks that allow to tackling effectively the complexrelationships of large multi-component systems.File | Dimensione | Formato | |
---|---|---|---|
Chapter 9_Bruneo_Quantitative Assessments of Distributed Systems.pdf
solo gestori archivio
Tipologia:
Versione Editoriale (PDF)
Licenza:
Non specificato
Dimensione
1.44 MB
Formato
Adobe PDF
|
1.44 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.