Βιβλιοθήκη ΟΠΑ - Ψηφιακό Αποθετήριο

ΠΥΞΙΔΑ Ιδρυματικό Αποθετήριο
και Ψηφιακή Βιβλιοθήκη

Όνομα χρήστη
Κωδικός πρόσβασης

Συλλογές :	Ιδρυματικό Αποθετήριο ΟΠΑ / AUEB Institutional Repository Σχολή Επιστημών και Τεχνολογίας της Πληροφορίας / School of Informatics Τμήμα Πληροφορικής / Department of Informatics Μεταπτυχιακές Εργασίες / Postgraduate dissertations

Τίτλος :	A time-based risk analysis of cascading failures

Δημιουργός :	Λύκου, Γεωργία Lykou, Georgia

Συντελεστής :	Gritzalis, Dimitrios (Επιβλέπων καθηγητής) Athens School of Economics and Business, Department of Informatics (Degree granting institution)

Τύπος :	Text

Φυσική περιγραφή :	95p.

Γλώσσα :	en

Περίληψη :	One of the most challenging problems in Critical Infrastructure (CI) protection is the assessment and mitigation of cascading failures across infrastructures. Dependency analysis is a computationally intensive problem and various models worldwide have been examined in this work, which have been proposed for evaluating potential cascading effects and cumulative security risk due to high-order dependencies between CIs. The problem intensifies when attempting a dynamic time-based dependency analysis. Further research is needed to examine how failure duration and contingency plans restoration time, can affect CIs protection and resilience. This thesis presents a time based extension of previous C.I. cascading effects and common-cause failure Risk Analysis models. The method assesses the risk arising from cascading failures, triggered by major or concurrent common-cause events, when Impact is evaluated by using a time-related, functional analysis, which takes into account the type of vulnerability and time performance of contingency plans. We employ different growth models, to capture slow, linear and fast evolving effects, but instead of using static projections, the evolution of each dependency is “objectified” by a fuzzy control system, which also considers the effect of near dependencies. To achieve this, the impact (and eventually, risk) of dependency is quantified on a time axis, into a form of many-valued logic. Furthermore, the methodology is extended by analyzing major failures triggered by common-cause cascading events, with the use of CIDA (Critical Infrastructure Dependency Analysis tool), which implements this extended risk-based methodology. CIDA aims to support decision makers to proactively analyze dynamic and complex dependency risk paths in two ways: (a) to identify potentially hidden dependencies of high impact and risk, before they are actually realized and (b) to simulate the effectiveness of alternative mitigation controls with different reaction time. New methodology results provided a sound and more accurate dynamic cascading risk analysis of interdependent CIs, showing that the use of time –based impact ranks is congruent with what is happening in real world outages or failures in CI operation.

Περίληψη :

One of the most challenging problems in Critical Infrastructure (CI) protection is the assessment and mitigation of cascading failures across infrastructures. Dependency analysis is a computationally intensive problem and various models worldwide have been examined in this work, which have been proposed for evaluating potential cascading effects and cumulative security risk due to high-order dependencies between CIs. The problem intensifies when attempting a dynamic time-based dependency analysis. Further research is needed to examine how failure duration and contingency plans restoration time, can affect CIs protection and resilience. This thesis presents a time based extension of previous C.I. cascading effects and common-cause failure Risk Analysis models. The method assesses the risk arising from cascading failures, triggered by major or concurrent common-cause events, when Impact is evaluated by using a time-related, functional analysis, which takes into account the type of vulnerability and time performance of contingency plans. We employ different growth models, to capture slow, linear and fast evolving effects, but instead of using static projections, the evolution of each dependency is “objectified” by a fuzzy control system, which also considers the effect of near dependencies. To achieve this, the impact (and eventually, risk) of dependency is quantified on a time axis, into a form of many-valued logic. Furthermore, the methodology is extended by analyzing major failures triggered by common-cause cascading events, with the use of CIDA (Critical Infrastructure Dependency Analysis tool), which implements this extended risk-based methodology. CIDA aims to support decision makers to proactively analyze dynamic and complex dependency risk paths in two ways: (a) to identify potentially hidden dependencies of high impact and risk, before they are actually realized and (b) to simulate the effectiveness of alternative mitigation controls with different reaction time. New methodology results provided a sound and more accurate dynamic cascading risk analysis of interdependent CIs, showing that the use of time –based impact ranks is congruent with what is happening in real world outages or failures in CI operation.

Λέξη κλειδί :	Risk assessment Risk management Information systems Computer networks Εκτίμηση κινδύνου Διαχείριση κινδύνου Πληροφοριακά συστήματα Δίκτυα υπολογιστών

Ημερομηνία έκδοσης :	31-12-2014

Άδεια χρήσης :

Αρχείο: Lykou_2014.pdf

Τύπος: application/pdf

Είσοδος