Combining Markov and Semi-Markov Modelling for Assessing Availability and Cybersecurity of Cloud and IoT Systems

Abstract

This paper suggests a strategy (C5) for assessing cloud and IoT system (CIS) dependability, availability, and cybersecurity based on the continuous collection, comparison, choice, and combination of Markov and semi-Markov models (MMs and SMMs). It proposes the systematic building of an adequate and accurate model to evaluate CISs considering (1) continuous evolution of the model(s) together with systems induced by changes in the CIS or physical and cyber environment parameters; (2) the necessity of collecting data on faults, failures, vulnerabilities, cyber-attacks, privacy violations, and patches to obtain actual data for assessment; (3) renewing the model set based on analysis of CIS operation; (4) the possibility of choice and utilizing “off-the-shelf” models with understandable techniques for their development to assure improved accuracy of assessment; (5) renewing the models during application of CIS by time, component or mixed combining, taking into consideration different operation and maintenance events. The results obtained were algorithms for data collection and analysis, choice, and combining appropriate MM and SMMs and their different types, such as multi-fragmental and multiphase models, considering changing failure rates, cyber-attack parameters, periodical maintenance, etc. To provide and verify the approach, several private and public clouds and IoT systems were researched and discussed in the context of C5 and proposed algorithms.

Security-Informed Safety Analysis of Autonomous Transport Systems Considering AI-Powered Cyberattacks and Protection

“…Future research directions are as follows: enhancing models and techniques combining AIQM and SISMECA approaches, including refinement of AI quality and various XMECA templates models to minimise uncertainties and dependencies on expert errors; quantitative assessment of cybersecurity and safety evaluations entropy in the context of attributes decomposition on the application of AIQM and SISMECA; completing tool-based support of AIQM&SISMECA-based techniques for different actors (regulators, developers, operators, and customers) considering their specific responsibility and scenarios of AI-powered protection against AI-powered attacks; developing multi-criteria techniques for choosing countermeasures considering security, reliability, and safety attributes at component and system levels; combining the various assessment methods based on Markov and semi-Markov models to assess embedded and distributed IoT/cloud systems for ATSs [ 53 , 54 , 55 ]. …”

“…combining the various assessment methods based on Markov and semi-Markov models to assess embedded and distributed IoT/cloud systems for ATSs [ 53 , 54 , 55 ].…”

Multi-fragmental and multi-phase availability models of the safety-critical I&C systems with two-cascade redundancy

“…It should be emphasized that the investigation of models of such size is a challenging task. In studies [16], [17], [18], the generalization of individual blocks was performed, as well as the detailing of the system operation within specific fragments. The developed macro model provides the following advantages:…”

“…This work considers a macro model for analyzing the operation of the reactor trip system (RTS) of nuclear power plants (NPP) and other safety-critical systems taking into account various types of failures, in particular errors in the diagnostic units of the main and diverse subsystems. This paper is a continuation of scientific research that was described in studies [16], [17], [18]. Publications [16], [17] have been dedicated to the research of multi-cascade redundant control systems considering failures of supervision units, software versions, and recovery processes.…”

“…Publications [16], [17] have been dedicated to the research of multi-cascade redundant control systems considering failures of supervision units, software versions, and recovery processes. An approach based on combining Markov and semi-Markov modeling for assessing the availability and safety of IoT and cloud-based systems considering changes in requirements and environment parameters was researched in [18].…”