Fundamental Concepts of Cyber Resilience: Introduction and Overview

Motivation: Why Cyber Resilience?

• Threats affect many cyber functions, from denial of service and data theft to code changes and malware infection.Society relies on interconnected cyber systems across everyday activities, finance, defense, and air traffic.Source: Motivation: Why Cyber Resilience? · p. 1
• Risk assessment struggles to account for threats and vulnerabilities spanning interdependent systems and the cascading effects of outages or losses.The problem is especially acute for highly complex, interconnected systems.Source: Motivation: Why Cyber Resilience? · pp. 1–2
• Unpredictable, rapidly changing threats make hardening against identified threats impossible without effectively disconnecting systems from the internet.Cyber systems therefore need to adapt to threats and bounce back from attacks on vital functions.Source: Motivation: Why Cyber Resilience? · p. 2
• Cyber resilience means a system’s ability to prepare, absorb, recover, and adapt to adverse effects, especially cyberattacks.The term can also refer to system features or components that enable resilience.Source: Motivation: Why Cyber Resilience? · p. 2

Resilience and Systems

• Cyber resilience considers interconnected hardware, software, and sensing components to support system recovery, sustained operations, and mission execution.It bridges sustaining system operations with carrying out the mission.Source: Resilience and Systems · p. 2, Figure 1 · p. 3
• Resilience draws on ecological, social, psychological, organizational, and engineering perspectives, whose concepts and methods remain weakly linked.Different technical languages also make communication across disciplines and applications difficult.Source: Resilience and Systems · p. 3
• Common resilience features include critical functions, thresholds, cross-scale interactions, and memory and adaptive management.These features are related to the temporal phases of the National Academies definition of resilience.Source: Resilience and Systems · pp. 3–4, Table 1
• Critical functions identify valued system performance, while thresholds mark stress levels beyond which a regime shift may occur.The threshold concept is linked to whether a system can absorb a shock.Source: Resilience and Systems · pp. 4, 6
• Recovery time measures the duration of degraded performance, while memory and adaptive management capture learning and management responses to disruption.Adaptive management uses learning from prior events to change management or other responses.Source: Resilience and Systems · pp. 4, 6

Resilience and Related Properties of Systems

• Resilience compares systems with equal initial performance and degradation by asking which recovers to a higher performance level after a given period.Assessment assumes the system has been affected and focuses on recovery speed.Source: Resilience and Related Properties of… · pp. 7–8
• Risk combines the likelihood of an undesirable event with its impact, but does not describe how well a system absorbs an attack or recovers.Resilience assessment and management also address residual risk and responses to unknown or emerging threats.Source: Resilience and Related Properties of… · pp. 8–9
• Robustness measures performance maintained under unexpected impacts, whereas resilience concerns how quickly the system recovers from resulting damage.A system that lacks robustness may be damaged beyond recovery, so both properties must be understood together.Source: Resilience and Related Properties of… · p. 9

Making Business Case for Cyber Resilience

• Uncertain threat intensity, frequency, vulnerabilities, and consequences make risk calculations less certain and can produce costly mitigation plans.Hypothetical threats may imply many vulnerabilities and catastrophic failures that are unaffordable to mitigate, absorb, or recover from.Source: Making Business Case for Cyber… · p. 10
• Stakeholders may prefer some performance loss in one system part over degradation in another, so spending can fail to match their values.Misaligned spending can leave stakeholders dissatisfied despite the expense, or lead to litigation that interferes with risk reduction.Source: Making Business Case for Cyber… · p. 10
• Hardening all components against unknown threats becomes increasingly expensive, contributing to investment stagnation and leaving infrastructure unprepared.Interdependency among systems is also difficult to characterize and quantify.Source: Making Business Case for Cyber… · p. 10, Figure 1 · p. 10
• Accepting current risk levels rather than demanding greater preventive measures can free funding for resilience enhancement.The chapter contrasts accepting risk level (a) with risk level (b) as an example of the reallocation.Source: Making Business Case for Cyber… · pp. 10–11, Figure 3 · p. 11
• The academic community is urged to develop decision models for balancing investment in risk reduction against resilience and recovery improvement.The proposed models would identify optimal investment across these options.Source: Making Business Case for Cyber… · p. 11

Assessment of Cyber Resilience

• Resilience assessment tools range from checklists to visualizations and custom network models, with varied outputs that can be difficult for users to choose among.Outputs include maps, scores, and process-time graphs.Source: Assessment of Cyber Resilience · p. 11
• Metric-based approaches assess system performance through component or function measures, whereas model-based approaches use system configuration and scenario analysis to predict system evolution.Metrics are measurable system properties that quantify the degree to which objectives are achieved.Source: Assessment of Cyber Resilience · pp. 11–12
• A lack of universally applicable resilience metrics and difficulty formalizing relevant value systems have hindered wide implementation of metric-based methods.Decision analysis and social and economic valuation can help assess trade-offs between resilience attributes and factors such as cost, environmental impact, and risk reduction.Source: Assessment of Cyber Resilience · p. 12
• Model-based approaches require knowledge of critical functions, mission, temporal patterns, thresholds, memory, and adaptation, while many advanced models are limited by data-intensive requirements.Process models are difficult to construct and information-consuming; statistical approaches require substantial system-performance data.Source: Assessment of Cyber Resilience · p. 12
• The tiered approach moves from broad function framing to simple system relationships and then parameterized models, stopping when decision makers have enough information for actionable investments.Tier I uses existing data, expert judgment, and conceptual models; Tier II can reveal dependencies behind cascading failures.Source: Assessment of Cyber Resilience · p. 14

Approaches to Improving Cyber Resilience

• Hidden links across physical, communication, information, social, and cognitive networks make failure paths difficult to anticipate, and greater complexity can both aid restoration and conceal cascading risks.The authors advise avoiding added complexity unless it directly supports resilience functions.Source: Approaches to Improving Cyber Resilience · pp. 14–15
• Scale-free networks are more robust to random node failures than exponential-degree networks, but more vulnerable to targeted removal of high-degree nodes.Source: Approaches to Improving Cyber Resilience · p. 15
• Additional capacity, storage, and parallel functions can improve resilience, but added resources and complexity may make recovery harder after unexpected failures.Identical redundant hardware or software can be compromised by the same malware, while diverse redundancy increases complexity.Source: Approaches to Improving Cyber Resilience · pp. 15–16
• Reversible designs should allow compromised components to enter a safe mode and recover, but conventional fail-safe disconnection may undermine resilience when a component supports damage absorption.Source: Approaches to Improving Cyber Resilience · p. 16
• Designing links to limit failure propagation can reduce cascading damage, while local buffers let data or commodity networks continue serving users when original sources become unreachable.Source: Approaches to Improving Cyber Resilience · pp. 16–17
• Prepared human or artificial agents can absorb, recover, and adapt; versatile, interchangeable agents support reconfiguration when resources are lost.Human agents need training, roles, resources, and processes; agent capacity describes how many functions can be performed simultaneously.Source: Approaches to Improving Cyber Resilience · pp. 17–18
• Resilience measures should account for adversary capabilities and tactics, because adversaries may target absorption and recovery mechanisms and agents.Artificial agents may need controlled destructive actions and stealth to reduce the risk of detection.Source: Approaches to Improving Cyber Resilience · pp. 17–18
• Proposed resilience measures require comparative, high-fidelity analysis that captures propagation, interactions, feedback, higher-order effects, and adversary actions.Source: Approaches to Improving Cyber Resilience · p. 18

• One chapter frames cyber resilience through system performance, while another uses mission risk to support measurable assessment and comparison tailored to specific systems.Source: Preview of the Book · p. 19
• Assessment chapters offer framework-based best practices and dependency methods, including analysis of direct and indirect effects of dependency loss.One dependency method includes backup recovery and assessment over time; the other omits redundancy and delayed impacts.Source: Preview of the Book · pp. 19–20
• Percolation theory measures network resilience through connectivity after node or edge failures and adds measures of adaptability and recoverability alongside fault tolerance.Source: Preview of the Book · p. 20
• The book previews analytical cyberattack models built with stochastic networks, multidimensional organizational simulation, and a simulation integration toolkit illustrated with transportation networks.Source: Preview of the Book · pp. 21, 25
• Enhancement approaches include systems-engineering frameworks, contingency planning, diversity, redundancy, analytic monitoring, coordinated defense, segmentation, and active defense.Active defense combines automated and human-directed techniques; the overview notes that widely available commercial solutions are lacking.Source: Preview of the Book · pp. 21–22
• Human factors chapters address the role of people in cascading vulnerability and insider threats, including how to balance employee rules between too many and too few.Source: Preview of the Book · pp. 22–23
• The book surveys biologically inspired AI techniques and evaluates resilience tactics economically, including estimates of mitigation costs and effectiveness.The economic discussion uses a benefit-cost analysis framework and connects quicker post-disaster recovery with prospects for reducing business interruption.Source: Preview of the Book · p. 23
• Illustrative applications span regional infrastructure, IoT and smart cities, dynamically randomized resources for smart-city services, and supply-chain resilience.The cases are presented as examples that may inform approaches to other system classes; the book also reviews research questions still to be addressed.Source: Preview of the Book · pp. 24–25