TY - JOUR
T1 - A Collaborative Approach for Minimal-Cost Monitor Deployment in Cloud Environment
AU - Tung, Yuan Hsin
AU - Tseng, Shian Shyong
AU - Tsai, Wei Tek
N1 - Funding Information:
This work was partially supported by National Science Council of the Republic of China under contracts NSC101-2511-S-468-007-MY3 and NSC101-2511-S-468-001.
Publisher Copyright:
© 2015 World Scientific Publishing Company.
PY - 2015/8/1
Y1 - 2015/8/1
N2 - Monitoring is widely applied in problem diagnosis, fault localization, and system maintenance. And since the cloud infrastructure is complex, the applications on the cloud are therefore complex, which makes monitoring in cloud more difficult. Rich monitors that contain composite and heterogeneous probes are often used in service-oriented system monitoring. These rich monitors often involve multiple entities, and the interpretation may require expert opinions from multiple domains. This paper proposes a knowledge-based collaborative monitoring approach to find out minimal cost monitor deployment in a cloud environment. The approach contains two main phases. In the knowledge acquisition phase, three acquisition tables, monitor-probe relationship matrix, cost of monitoring, and probe-problem dependence matrix, are generated according to diagnosis ontology and monitor ontology acquired from domain experts. And then based upon the three acquisition tables and three consensus building strategies, we formulate the problem of optimizing the cost of monitoring as an Integer Linear Programming (ILP) problem, which is NP-Complete. In the monitor deployment phase, the proposed algorithm applies two heuristic rules to address the problem. Three experiments are conducted to evaluate the performance of the proposed approach. The results from the experiments show that our approach is effective and produce quality approximate solutions in monitor deployment.
AB - Monitoring is widely applied in problem diagnosis, fault localization, and system maintenance. And since the cloud infrastructure is complex, the applications on the cloud are therefore complex, which makes monitoring in cloud more difficult. Rich monitors that contain composite and heterogeneous probes are often used in service-oriented system monitoring. These rich monitors often involve multiple entities, and the interpretation may require expert opinions from multiple domains. This paper proposes a knowledge-based collaborative monitoring approach to find out minimal cost monitor deployment in a cloud environment. The approach contains two main phases. In the knowledge acquisition phase, three acquisition tables, monitor-probe relationship matrix, cost of monitoring, and probe-problem dependence matrix, are generated according to diagnosis ontology and monitor ontology acquired from domain experts. And then based upon the three acquisition tables and three consensus building strategies, we formulate the problem of optimizing the cost of monitoring as an Integer Linear Programming (ILP) problem, which is NP-Complete. In the monitor deployment phase, the proposed algorithm applies two heuristic rules to address the problem. Three experiments are conducted to evaluate the performance of the proposed approach. The results from the experiments show that our approach is effective and produce quality approximate solutions in monitor deployment.
KW - Monitoring
KW - cloud computing
KW - collaborative knowledge acquisition
KW - ontology-based
KW - rich monitor
UR - http://www.scopus.com/inward/record.url?scp=84945975960&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84945975960&partnerID=8YFLogxK
U2 - 10.1142/S0218194015500126
DO - 10.1142/S0218194015500126
M3 - Article
AN - SCOPUS:84945975960
SN - 0218-1940
VL - 25
SP - 935
EP - 960
JO - International Journal of Software Engineering and Knowledge Engineering
JF - International Journal of Software Engineering and Knowledge Engineering
IS - 6
ER -