[Paper Review] A Survey on Resilience in the IoT: Taxonomy, Classification and Discussion of Resilience Mechanisms

Christian Berger, Philipp Eichhammer+4 more|arXiv (Cornell University)|Sep 6, 2021

Software System Performance and ReliabilityComputer Science155 references73 citations

TL;DR

This paper presents a comprehensive taxonomy and classification of resilience mechanisms in IoT systems, analyzing redundancy, monitoring, protection, and recovery techniques. It evaluates their practical applicability across multi-layered IoT architectures, emphasizing cross-layer resilience, self-adaptation, and trade-offs in implementation, offering a systematic guide for developers and architects to design dependable, secure, and adaptive IoT ecosystems.

ABSTRACT

Internet-of-Things (IoT) ecosystems tend to grow both in scale and complexity as they consist of a variety of heterogeneous devices, which span over multiple architectural IoT layers (e.g., cloud, edge, sensors). Further, IoT systems increasingly demand the resilient operability of services as they become part of critical infrastructures. This leads to a broad variety of research works that aim to increase the resilience of these systems. In this paper, we create a systematization of knowledge about existing scientific efforts of making IoT systems resilient. In particular, we first discuss the taxonomy and classification of resilience and resilience mechanisms and subsequently survey state-of-the-art resilience mechanisms that have been proposed by research work and are applicable to IoT. As part of the survey, we also discuss questions that focus on the practical aspects of resilience, e.g., which constraints resilience mechanisms impose on developers when designing resilient systems by incorporating a specific mechanism into IoT systems.

Motivation & Objective

• To establish a unified understanding of resilience in IoT by analyzing definitions, properties, and classifications across academic literature.
• To identify and categorize key resilience mechanisms—redundancy, monitoring, protection, and recovery—applicable to heterogeneous, multi-layered IoT systems.
• To examine practical constraints imposed by resilience mechanisms on developers and runtime platforms during system design and deployment.
• To address the evolving needs of IoT systems, including scalability, heterogeneity, dynamic reconfiguration, and cross-domain administrative views.
• To extend traditional fault-tolerance models by incorporating a continuum of resilience states, including 'best-effort' recovery and degraded functionality under persistent faults.

Proposed method

• Proposes a taxonomy of resilience based on system properties such as availability, confidentiality, integrity, reliability, and safety.
• Classifies resilience mechanisms into four core categories: redundancy (e.g., replication, backup routing), monitoring (e.g., anomaly detection, health checks), protection (e.g., access control, intrusion detection), and recovery (e.g., self-healing, reconfiguration).
• Analyzes mechanisms through a layered IoT architecture (device, edge, cloud), evaluating cross-layer applicability and integration challenges.
• Evaluates mechanisms using real-world IoT use cases, including e-Health, smart cities, and industrial IoT, with examples from Kafka, Kubernetes, and blockchain-based middleware.
• Introduces a continuum model of resilience, moving beyond binary fault tolerance to include partial recovery and degraded operation under sustained faults.
• Reviews pattern-based and middleware-driven approaches (e.g., Resilience Manager, shadow devices, containerization) to enable modular and composable resilience.

Experimental results

Research questions

• RQ1 What is a consistent and comprehensive taxonomy of resilience in IoT systems, and how does it differ from traditional dependability models?
• RQ2 How do resilience mechanisms (redundancy, monitoring, protection, recovery) function across the layered architecture of IoT systems (device, edge, cloud)?
• RQ3 What practical constraints do resilience mechanisms impose on developers during system design and on execution platforms at runtime?
• RQ4 How can resilience be quantified and measured in IoT systems, and what metrics best reflect application-specific resilience requirements?
• RQ5 To what extent can modern IoT systems achieve self-adaptation and 'best-effort' resilience beyond traditional fault tolerance?

Key findings

• Resilience in IoT extends beyond fault tolerance to include security, privacy, and dynamic adaptability, requiring a holistic approach that integrates dependability and cyber-physical system properties.
• Redundancy mechanisms such as data replication via Apache Kafka and service replication through Kubernetes enable high availability and fault tolerance in edge and cloud layers.
• Monitoring and anomaly detection techniques, including health pattern sequencing and distributed logging, are critical for early fault and attack detection in heterogeneous IoT environments.
• Protection mechanisms like access control, authorization, and intrusion detection are essential for securing open, large-scale IoT systems with evolving trust boundaries.
• Recovery mechanisms such as automatic reconfiguration, shadow devices, and container-based fault tolerance allow systems to restore functionality after disruptions, even under degraded conditions.
• The integration of blockchain and smart contracts enables trustless, resilient service coordination in decentralized IoT marketplaces, though with performance trade-offs.