TY - GEN
T1 - Failure localization in transparent optical networks
AU - Staessens, Dimitri
AU - Manousakis, Konstantinos
AU - Colle, Didier
AU - Mahlab, Uri
AU - Mahlab, Uri
AU - Pickavet, Mario
AU - Varvarigos, Emmanouel
AU - Demeester, Piet
PY - 2010/12/1
Y1 - 2010/12/1
N2 - In this paper, we address failure localization from both a practical and a theoretical perspective. After summarizing the state-of-the-art of failure localization algorithms and monitoring techniques, an overview of the most prevalent failures in optical core networks is presented. We review the role of the Optical Supervisory Channel and how it reports problems to the management plane.We analyze different equipment, investigating where most failures occur and how these failures can be monitored. We conclude that in-band OSNR monitoring is the most important monitoring technique for failure localization purposes. We give a general probabilistic model for failure localization and assess its limitations using the mutual information metric. We give a simple example for computing this mutual information and show that is it a valid metric for evaluation of the failure localization problem. For practical applications, with imperfect monitoring equipment and countless possible failures, the mutual information may be prohibitingly low. Initial analysis of the problem shows that we need intense and accurate monitoring in order to increase the mutual information for the problem and to be able to localize failures accurately.
AB - In this paper, we address failure localization from both a practical and a theoretical perspective. After summarizing the state-of-the-art of failure localization algorithms and monitoring techniques, an overview of the most prevalent failures in optical core networks is presented. We review the role of the Optical Supervisory Channel and how it reports problems to the management plane.We analyze different equipment, investigating where most failures occur and how these failures can be monitored. We conclude that in-band OSNR monitoring is the most important monitoring technique for failure localization purposes. We give a general probabilistic model for failure localization and assess its limitations using the mutual information metric. We give a simple example for computing this mutual information and show that is it a valid metric for evaluation of the failure localization problem. For practical applications, with imperfect monitoring equipment and countless possible failures, the mutual information may be prohibitingly low. Initial analysis of the problem shows that we need intense and accurate monitoring in order to increase the mutual information for the problem and to be able to localize failures accurately.
UR - http://www.scopus.com/inward/record.url?scp=79951474421&partnerID=8YFLogxK
U2 - 10.1109/ICUMT.2010.5676577
DO - 10.1109/ICUMT.2010.5676577
M3 - Conference Paper
AN - SCOPUS:79951474421
SP - 589
EP - 594
BT - 2010 International Congress on Ultra Modern Telecommunications and Control Systems and Workshops, ICUMT 2010
ER -