Workload-sensitive Timing Behavior Anomaly Detection in Large Software Systems.

van Hoorn, Andre (2007) Workload-sensitive Timing Behavior Anomaly Detection in Large Software Systems. (Diploma thesis), University of Oldenburg, Germany, Oldenburg, Germany, 125 pp.

[thumbnail of vanhoorn2007WorkloadSensitiveAnomalyDetectionInLargeSoftwareSystems.pdf]
Preview
Text
vanhoorn2007WorkloadSensitiveAnomalyDetectionInLargeSoftwareSystems.pdf - Published Version

Download (4MB) | Preview
[thumbnail of vanhoorn2007WorkloadSensitiveAnomalyDetectionInLargeSoftwareSystems-slides-2x3.pdf]
Preview
Text
vanhoorn2007WorkloadSensitiveAnomalyDetectionInLargeSoftwareSystems-slides-2x3.pdf - Presentation

Download (2MB) | Preview

Abstract

Anomaly detection is used for failure detection and diagnosis in large software systems to reduce repair time, thus increasing availability. A common approach is building a model of a system's normal behavior in terms of monitored parameters, and comparing this model with a dynamically generated model of the respective current behavior. Deviations are considered anomalies, indicating failures. Most anomaly detection approaches do not explicitly consider varying workload. Assuming that varying workload leads to varying response times of services provided by internal software components, our hypothesis is as follows: a novel workload-sensitive anomaly detection is realizable, using statistics of workload-dependent service response times as a model of the normal behavior. The goals of this work are divided into three parts. First, an application-generic technique will be developed to model and generate realistic workload based on an analytical workload model notation to be specified. Applying this technique to a sample application in a case study, the relation between workload intensity and response times will be statistically analyzed. Based on this, a workload-sensitive anomaly detection prototype will be implemented.

Document Type: Thesis (Diploma thesis)
Keywords:
Research affiliation: Kiel University > Software Engineering
Open Access Journal?: Yes
Date Deposited: 04 Oct 2012 06:22
Last Modified: 11 Oct 2012 14:53
URI: https://oceanrep.geomar.de/id/eprint/15491

Actions (login required)

View Item View Item