Home // INTERNET 2017, The Ninth International Conference on Evolving Internet // View article

Verification of Openstack Operation with Normality-degree of Workflows

Authors:
Ryota Mizutani
Takeshi Usui
Yoshinori Kitatsuji

Keywords: NFV; Openstack; Log message; Normality-degree; Similarity-degree;

Abstract:
Verifying job executions in operating virtual networks has emerged as a crucial issue for network operators. Now, tracking and monitoring log messages of the operations is a typical approach. However, a monitoring system to detect a critical operation is based on text matching with the previously obtained log messages indicating normal operations. It basically has a limitation that similar but different log message flows become indeterminable to judge the normality of the operation. We propose a solution that estimates the normality-degree for the indeterminable log message flows. We evaluate the difference in normality-degrees between normal and abnormal completion message flows. We show that the proposed method distinguishes these cases more clearly than the conventional method. This indicates that the proposed method enables the system administrator to classify job execution into three types (i.e., normal, semi-normal, and abnormal completions). As a result, the proposed method allows the system administrator to focus on critical job executions in troubleshooting operations.

Pages: 7 to 13

Copyright: Copyright (c) IARIA, 2017

Publication date: July 23, 2017

Published in: conference

ISSN: 2308-443X

ISBN: 978-1-61208-574-6

Location: Nice, France

Dates: from July 23, 2017 to July 27, 2017