On-line failure prediction in safety-critical systems

Baldoni, Roberto; Montanari, Luca; Rizzuto, Marco

doi:10.1016/j.future.2014.11.015

Cited by 39 publications

(34 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Considering that we are targeting safety-critical systems, we prefer to reduce the amount of missed detections (FN), even at the cost of a higher rate of FP. For this reason, we selected as reference metric the F-Score (2), which considers the recall more relevant than the precision: the F-Score(2) for each anomaly checker is computed, and checkers are selected accordingly (choosing the n best, or those whose F-Score(2) > δ).…”

Section: Anomaly Checkers For Each Indicator We Build Three Types Omentioning

confidence: 99%

“…Anomaly detectors have been proposed as error detectors [10] or failure predictors [2], based on the hypothesis that the activation of a fault (for error detection) or an error (for failure prediction) manifests as increasingly unstable performance-related behavior before escalating into a failure. The anomaly detector is in charge to observe these fluctuations providing a response to the administrator as soon as it can, triggering proactive recovery or dumping critical data.…”

Section: State Of the Art And Comparison With Other Solutionsmentioning

confidence: 99%

“…The anomaly detector is in charge to observe these fluctuations providing a response to the administrator as soon as it can, triggering proactive recovery or dumping critical data. Reviewing state of the art it is possible to notice that the most used layers are the network [2], [3] and the operating system [6], [11]. This is not surprising since most of the systems include these layers: building solutions which fetch data from these layers allow building frameworks that fit in a very wide range of contexts.…”

Section: State Of the Art And Comparison With Other Solutionsmentioning

confidence: 99%

“…A detailed overview of anomaly detection frameworks can be found in [1]. Here we focus on three anomaly detection frameworks [2], [3], [14] addressing error detection/failure prediction where the authors reported the measurements for detection accuracy metrics (i.e., precision and recall). They observe indicators from multiple layers as the framework presented here does.…”

Section: State Of the Art And Comparison With Other Solutionsmentioning

confidence: 99%

“…Summarizing, our main findings are: i) describing how context-awareness on the SOA services can be used to improve detection; ii) defining a methodology and the associated framework for anomaly detection in dynamic contexts using contextawareness; iii) structuring a multi-layer anomaly detection module observing operating system, middleware (Java Virtual Machine, JVM) and network layers, iv) assessing the whole solution on a case study, showing the obtained detection accuracy, which is presented using well-known metrics and v) compare our detection system with state-of-the-art [2], [3], [14] solutions exercised in less dynamic contexts.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Context-Awareness to Improve Anomaly Detection in Dynamic Service Oriented Architectures

Zoppi

Ceccarelli

Bondavalli

2016

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. Revealing anomalies to support error detection in softwareintensive systems is a promising approach when traditional detection mechanisms are considered inadequate or not applicable. The core of anomaly detection lies in the definition of the expected behavior of the observed system. Unfortunately, the behavior of complex and dynamic systems is particularly difficult to understand. To improve the accuracy of anomaly detection in such systems, in this paper we present a contextaware anomaly detection framework which acquires information on the running services to calibrate the anomaly detection. To cope with system dynamicity, our framework avoids instrumenting probes into the application layer of the observed system monitoring multiple underlying layers instead. Experimental evaluation shows that the detection accuracy is increased considerably through context-awareness and multiple layers monitoring. Results are compared to state-of-the-art anomaly detectors exercised in demanding more static contexts.

show abstract

Section: Anomaly Checkers For Each Indicator We Build Three Types Omentioning

confidence: 99%

Section: State Of the Art And Comparison With Other Solutionsmentioning

confidence: 99%

Section: State Of the Art And Comparison With Other Solutionsmentioning

confidence: 99%

Section: State Of the Art And Comparison With Other Solutionsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Context-Awareness to Improve Anomaly Detection in Dynamic Service Oriented Architectures

Zoppi

Ceccarelli

Bondavalli

2016

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

Predicting locally manageable resource failures of high availability clusters

Somasekaram

Călinescu

2022

Softw Pract Exp

View full text Add to dashboard Cite

Critical services from domains as diverse as finance, manufacturing and healthcare are often delivered by complex enterprise applications (EAs). High-availability clusters (HACs) are software-managed IT infrastructures that enable these EAs to operate with minimum downtime. This paper presents a novel Bayesian decision network model to improve the failure detection capabilities of the HACs components using a comprehensive set of characteristics for the analysed component. The model then combines these characteristics to predict whether the failure of this component can be managed locally at the failed component level without propagating the failure to upper-level components and causing a complete system failure. By improving the detection capabilities and predicting locally manageable failures, the model improves the decision-making process of HACs, and has the potential to reduce the downtime and improve availability for the applications protected by HACs. The model uses the capabilities of the Bayesian decision networks, which combines Bayesian networks with the utility theory, to assign weights to different characteristics and consolidate the related variables to output the result. The model evaluation in a realistic testbed environment with three servers, an established HAC and a well-known EA shows that the model can improve the area under the Receiver Operating Characteristic (ROC) curve for prediction of locally manageable failures by up to 9.05% compared to the baseline HAC results.

show abstract

A Taxonomy of Event Prediction Methods

Gmati

Chakhar²,

Chaari

et al. 2019

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Most of existing event prediction approaches consider event prediction problems within a specific application domain while event prediction is naturally a cross-disciplinary problem. This paper introduces a generic taxonomy of event prediction approaches. The proposed taxonomy, which oversteps the application domain, enables a better understanding of event prediction problems and allows conceiving and developing advanced and context-independent event prediction techniques.

show abstract

On-line failure prediction in safety-critical systems

Cited by 39 publications

References 29 publications

Context-Awareness to Improve Anomaly Detection in Dynamic Service Oriented Architectures

Context-Awareness to Improve Anomaly Detection in Dynamic Service Oriented Architectures

Predicting locally manageable resource failures of high availability clusters

A Taxonomy of Event Prediction Methods

Contact Info

Product

Resources

About