Metrics For Quantifying Interdependencies

Abstract

The quantification of interdependencies is a major challenge when attempting to analyze the behavior of critical infrastructures. This paper presents a taxonomy of interdependency quantification metrics based on their information content, decision support and risk analysis capabilities, and computational costs. The paper also discusses a systematic approach for computing metrics and performance indices that measure the effectiveness of strategies designed to enhance critical infrastructure protection and resilience. A case study is used to illustrate the computation of the metrics and performance indices, and their application to the analysis of critical infrastructure interdependencies.Keywords: Interdependencies, metrics, federated simulation IntroductionA critical infrastructure is a physical system that, if disrupted, can seriously affect the national security, economy and social welfare of a nation. Examples of critical infrastructures include telecommunications, electric power systems, natural gas and oil, banking and finance, transportation, water supply systems, government and emergency services [1]. Clearly, modern society cannot function if large portions of the critical infrastructure are disrupted or destroyed.In order to understand the behavior of critical infrastructures, it is necessary to accurately model and quantify their interdependencies [14] [7,9] and graph theory [15,16]. Quantitative approaches typically engage discrete simulation or agent-based modeling and simulation (ABMS) [3][4][5][6]14]. Please use the following format when citing this chapter:Casalicchio, E. and Galli, E., 2008, in IFIP International Federation for Information Processing, Volume 290; Critical Infrastructure Protection II, eds. Papa, M., Shenoi, S., (Boston: Springer), pp. 215-227. CRITICAL INFRASTRUCTURE PROTECTION IIWhile considerable research has focused on interdependency modeling and analysis, very few efforts have examined the issue of quantifying interdependencies. Zimmerman [17] has proposed explicit metrics for quantifying interdependencies. One metric measures the "direction" of infrastructure failures as the ratio between the number of times one type of infrastructure causes damage to another type of infrastructure and the number of times other types of infrastructures cause damage to the first type of infrastructure. Zimmerman and Restrepo [18] have specified a metric that measures the duration of cascading effects; they use this metric to quantify the effects of U.S. power grid outages on other infrastructures. This paper presents a taxonomy that classifies interdependency metrics on the basis of their information content, decision support and risk analysis capabilities, and computational costs. In addition, it describes systematic approaches for computing metrics using system or model observations, and for calculating performance indices that measure the effectiveness of strategies designed to enhance critical infrastructure protection and resilience.In general, interdependency metrics may be cla...

A methodology for modeling and measuring interdependencies of information and communications systems used for public administration and eGovernment services

“…Casalicchio and Galli [6] have presented a taxonomy of interdependency quantification metrics for critical infrastructures. The metrics are based on the information content, decision support and risk analysis capabilities, and computational costs.…”

Quantification of dependencies between electrical and information infrastructures

“…More recently, the issue of how to measure the level of interdependencies and to identify appropriate metrics for quantifying their strength is being addressed in a few studies, such as [29][30][31]. These studies are orthogonal to our objective and could be profitably employed to guide the analysis of EPS by concentrating on those interdependencies showing an higher degree of strength.…”

Risk-Based Criticality Analysis

“…Examples include critical infrastructure protection layers, the implications of dependencies between layers, and the multi-dimensional nature of the impact of an incident [1]. Approaches for interdependency identification, modeling, visualization and simulation should be embedded in risk analysis methodologies [5,11,19,23,24].…”