PaperlayerView exact revision on arXiv

Source-linked AI summary

Challenges and Opportunities in Securing the Industrial Internet of Things

Martin Serror, Sacha Hack, Martin Henze, Marko Schuba, Klaus Wehrle

arXiv:2111.11714v1cs.CRcs.NI

TL;DR

Industrial IoT security inherits consumer IoT problems while adding challenges from long-lived components, high connectivity, and safety-critical operations. The paper analyzes these differences and surveys countermeasures, finding that protection requires broad adaptation beyond device-level defenses.

  • Problem

    Industrial IoT security has distinct challenges because devices are long-lived, networks are highly connected, and operations may be safety- or mission-critical.

  • Method

    The paper compares consumer and industrial IoT security, derives IIoT-specific goals and challenges, and surveys research and best practices for applicability and security benefits.

  • Results

    The analysis concludes that a wide range of countermeasures is needed and that consumer-tailored measures must be adapted for industrial settings.

  • Takeaways & Limitations

    Legacy devices that cannot be patched require complementary network-layer mechanisms such as access policies and monitoring.

Abstract

from arXiv · show

Given the tremendous success of the Internet of Things in interconnecting consumer devices, we observe a natural trend to likewise interconnect devices in industrial settings, referred to as Industrial Internet of Things or Industry 4.0. While this coupling of industrial components provides many benefits, it also introduces serious security challenges. Although sharing many similarities with the consumer Internet of Things, securing the Industrial Internet of Things introduces its own challenges but also opportunities, mainly resulting from a longer lifetime of components and a larger scale of networks. In this paper, we identify the unique security goals and challenges of the Industrial Internet of Things, which, unlike consumer deployments, mainly follow from safety and productivity requirements. To address these security goals and challenges, we provide a comprehensive survey of research efforts to secure the Industrial Internet of Things, discuss their applicability, and analyze their security benefits.

I. INTRODUCTION

Industrial IoT security inherits consumer IoT vulnerabilities but introduces distinct challenges because industrial devices are longer-lived, networks are larger and more connected, and safety and productivity are central requirements. The paper surveys these challenges and maps relevant security approaches to IIoT applicability and benefits.

  • Motivation: Consumer IoT devices suffer severe security and privacy risks from missing features, poor updates, and insecure user configuration.These weaknesses have contributed to attacks against IoT devices.
  • Motivation: Successful attacks on industrial availability or operational safety can cause devastating effects extending beyond a single company.Examples include attacks on a German steel mill and the Ukraine power grid.
  • Distinct IIoT challenges: IIoT security differs from consumer IoT security because industrial devices have longer lifetimes, networks operate at larger scale, and device interconnections are more dynamic.These characteristics require prolonged patch management and complicate secure network design.
  • Contributions: The paper provides a comprehensive IIoT security survey that combines previously separate approaches and maps countermeasures to identified challenges.Its stated contributions include assessing vulnerabilities, risks, and threats in emerging IIoT deployments.
  • Approach: The paper uses consumer IoT security analysis to identify industrial differences and derive IIoT-specific security goals and challenges shaped by safety and productivity requirements.It then surveys research and best practices, discussing applicability and security benefits for IIoT scenarios.

II. IOT SECURITY THREATS

Consumer IoT security threats are commonly organized by perception, network, and application layers, with attacks and countermeasures associated with each layer. The figure presents this layered organization and its attack categories.

  • Layered attack taxonomy: IoT security surveys commonly classify attacks across perception, network, and application layers.The layered taxonomy supports assigning suitable countermeasures to the relevant architectural layer.
  • Perception layer: The perception layer covers sensing and actuation, with physical attacks targeting hardware and impersonation attacks exploiting missing authentication.Physical attackers may replace or damage components, while spoofing and Sybil attacks create false identities.
  • Figure overview: Fig. 1 links IoT reference layers with attack categories and countermeasures.Some attacks can occur across multiple layers.
  • Network layer: The network layer transports data between devices and applications through communication protocols and gateways, making it vulnerable to interception and routing disruption.Man-in-the-middle attacks can eavesdrop on or manipulate traffic, while sinkhole and selective-forward attacks threaten availability.

3) Application Layer:

The application layer delivers IoT services and can change independently through software, making it especially vulnerable to newly introduced weaknesses. Attacks include malicious code injection, data leakage, and denial of service.

  • 3) Application Layer:: Application-layer software may evolve without changes to lower layers, so software changes can introduce new vulnerabilities over time.This independence makes application-layer security especially challenging.
  • 3) Application Layer:: Malicious code injection exploits vulnerabilities to control infected devices and misuse their computing or communication resources.Reported motivations include recruiting devices for DDoS attacks and performing cryptojacking.
  • 3) Application Layer:: Data leakage uses software or service vulnerabilities to steal personal information, including through poorly configured cloud services.
  • 4) Multi-layer Attacks:: Denial-of-service attacks make resources or services unavailable, and centralized entities at any layer are especially vulnerable as potential single points of failure.Countermeasures should therefore account for the architectural layer where the attack occurs.
  • 3) Application Layer:: Consumer IoT security threats occur at every reference-model layer and require multiple security measures.

B. Countermeasures

IoT countermeasures span key management, authentication, access control, and privacy preservation, but IIoT deployments require mechanisms adapted to larger, more connected, and safety-critical environments. Consumer IoT methods provide a foundation while covering IIoT security only partially.

  • B. Countermeasures: A comprehensive IoT security taxonomy organizes countermeasures by key management, user and device authentication, access control, and privacy preservation.The taxonomy also relates these measures to IoT layers and attacks.
  • B. Countermeasures: Key management enables encrypted communication and authentication but must address deployment scalability and heterogeneity.Public-key mechanisms are easier to manage than symmetric schemes but require more computational resources.
  • B. Countermeasures: Authentication establishes user or device identity, while authorization and access control restrict access to required resources and services.Access-control rules and policies rely on authentication.
  • B. Countermeasures: Privacy preservation protects sensitive personal information through encryption and privacy-preserving data processing.Secure multiparty computation is given as an example of privacy-preserving processing.
  • C. Discussion: Consumer IoT security provides IIoT researchers with existing tools and lessons, but industrial deployments require new mechanisms beyond the device level.The paper concludes that a wide range of countermeasures must be adapted to industrial settings.

A. Common Features of Consumer IoT and Industrial IoT

Consumer and industrial IoT share pervasive connectivity and cloud integration, but IIoT deployments process more data locally and prioritize autonomous, safety- and mission-critical automation.

  • Consumer IoT devices mainly outsource processing to the cloud, whereas IIoT combines local processing with cloud services for optimization and new services.
  • Both domains increasingly connect pervasive computing devices to external cloud services as hardware and software costs decline.
  • IIoT strongly interconnects field devices, controllers, and servers, while increasing connectivity progressively merges industrial IT and OT networks.
  • Application: Consumer IoT assists humans, whereas IIoT augments automation through autonomous Machine-to-Machine connections for monitoring and control.
  • Application: IIoT includes safety- and mission-critical services, making security an important part of ensuring process safety.

2) Device:

IIoT deployments involve many more devices with much longer lifetimes than consumer deployments, while their sensing, monitoring, and control traffic is more regular and deterministic.

  • Device: Factories are expected to deploy an order of magnitude more connected devices than private homes.
  • Device: Industrial devices may operate for up to 30 years, compared with average consumer-device replacement after 3 to 5 years.
  • Device: IIoT security measures therefore need to be scalable, adaptable, and retrofittable.
  • Data Traffic: IIoT traffic mainly supports sensing, monitoring, and control, producing regular and deterministic flows that facilitate network policies and intrusion detection.
  • Because IIoT deployments differ in criticality, device lifetime, and traffic determinism, consumer-focused device security does not address their interdependence and tasks.

A. Requirements

IIoT security requirements prioritize availability and integrity because industrial services depend on safety and productivity. Its challenges include long-lived devices, scale, connectivity, critical processes, confidentiality, and human or insider failures.

  • Requirements: Availability and integrity are the foremost IIoT security requirements because industrial security is closely tied to safety and productivity.
  • Requirements: Availability loss can stop production or endanger human life, while manipulated sensor values can corrupt products, waste resources, or undermine safety.
  • Challenges: Long-lived legacy components require regular software updates and security mechanisms that address cumbersome update processes and increased breach risk.
  • Challenges: The large number of resource-constrained devices requires scalable, automatic deployment, configuration, operation, and security-management approaches.
  • Challenges: Strong IT–OT connectivity improves industrial efficiency and flexibility but makes isolation and unauthorized-access restriction more difficult, motivating network segmentation.
  • Challenges: Safety- and mission-critical processes require high availability and integrity, while security measures must avoid unacceptable communication and process latency.
  • Requirements: IIoT confidentiality protects customer data and business secrets while preserving authorized access for processing and analysis.
  • Challenges: Accidental failures, misconfigurations, and insider sabotage remain difficult to contain because networked components are increasingly interdependent.

V. SECURING THE IIOT

The paper surveys IIoT security approaches, explains how consumer-IoT methods can be adapted, and rates their coverage of identified challenges. Tailored cryptography must balance confidentiality and authentication against device constraints, key management, and latency.

  • The survey adapts consumer-IoT approaches to industrial contexts and rates each approach from no consideration to full consideration of the identified challenges.
  • Tailored Cryptography and Authentication: IIoT encryption supports confidentiality, authentication, and non-repudiation, but resource constraints favor lightweight symmetric-key cryptography.
  • Tailored Cryptography and Authentication: Symmetric-key methods create scalable key-management challenges, while public-key and symmetric-key cryptography can impose unacceptable delays in safety- and mission-critical processes.
  • Tailored Cryptography and Authentication: For periodic communication with static partners, partial precomputation can reduce latency and enable lightweight authentication and encryption.
  • The surveyed approaches include patch management, service isolation and access control, network monitoring and intrusion detection, and awareness, training, and assessment.
  • Tailored Cryptography and Authentication: Searchable public-key encryption protects cloud-outsourced data while allowing individual items to be retrieved and decrypted after search.

B. Patch Management

IIoT patch management must handle long-lived and legacy devices through automated, secure updates, extensive compatibility testing, vulnerability detection, and complementary hardware or network protections.

  • Patch deployment: Patching industrial systems requires extensive testing before installation to ensure compatibility, safety, and reduced process downtime.NIST recommends regression testing within systematic patch management.
  • Patch deployment: Automated firmware updates can simplify secure, standardized patching for resource-constrained IoT devices and can be adopted for IIoT deployments.The SUIT mechanism protects end-to-end firmware transmission and standardizes update descriptions.
  • Vulnerability detection: Vulnerability detection can test IIoT devices during idle times or analyze vulnerabilities using network graphs without impacting industrial processes.Idle-time testing is particularly useful for safety- and mission-critical processes.
  • Legacy protection: Zero-day vulnerabilities and discontinued vendor support leave some devices unpatchable, requiring defense-in-depth protections.Additional mechanisms are needed to protect legacy devices and prevent attacks from spreading to connected devices.
  • Hardware security: Trusted execution environments can securely execute safety- and mission-critical tasks on resource-constrained devices, but mainly apply to future hardware generations.ARM TrustZone reportedly processes data at least an order of magnitude faster than security controllers in the cited evaluation.

D. Network Monitoring and Intrusion Detection

IIoT intrusion detection must monitor traffic passively and at scale while accounting for real-time industrial processes, process behavior, human factors, and regular security assessment.

  • Network Monitoring and Intrusion Detection: Traditional IT intrusion detection systems cannot simply transfer to industrial networks dominated by real-time processes and operational constraints.Intrusion detection is especially important when preventive measures are inadequate or zero-day vulnerabilities are exploited.
  • Network Monitoring and Intrusion Detection: Passive, real-time-capable network monitoring can inspect communication traffic without interfering with critical industrial processes.Scalability also requires reducing the cost of packet inspection as device counts increase.
  • Network Monitoring and Intrusion Detection: Anomaly detection can reduce monitoring volume by an order of magnitude when traffic is monitored at flow level instead of packet level.The cited evaluation presents flow-level monitoring as a scalable IIoT solution.
  • Network Monitoring and Intrusion Detection: Process-aware IDSs combine network behavior with industrial-process models to distinguish attacks from system faults.The cited evaluation achieved 99.82% accuracy and could also identify some operator configuration errors.
  • Awareness, Training, and Assessment: Security protection also depends on workers, managers, policies, training, and regular assessment rather than technology alone.Assessment tools support evaluating security in larger IIoT deployments.

F. Summary and Recommendations

The survey complements prior IIoT research by mapping distinctive challenges to feasible countermeasures, especially for long-lived components and increased connectivity.

  • Related work: Prior IIoT surveys address challenges and requirements, increased connectivity, and the security impact of industrial processes.These streams include literature studies, cloud-based ICS security, data sharing, and process-oriented methodologies.
  • Related work: Process-oriented research highlights the strong coupling between safety and security and links vulnerabilities to industrial-process impacts.Some security measures can conflict with safety, making process context important for identifying challenges.
  • Contribution: This work extends prior surveys by discussing countermeasures and their feasibility, including retrofittable approaches for long-lived components.Its focus includes weaknesses arising from long-lived components and increased connectivity.
  • Contribution: The survey maps identified IIoT security challenges to corresponding countermeasures and combines previously independent approaches into a comprehensive survey.The mapping covers the unique challenges of IIoT deployments, particularly component longevity and connectivity.

VII. CONCLUSION

The paper concludes that IIoT security requires adapted, wide-ranging countermeasures that address long-lived devices, high connectivity, critical processes, insiders, and emerging technologies.

  • Conclusion: A wide range of countermeasures is needed, and consumer-oriented protections must be adapted for industrial settings.The paper identifies both substantial challenges and opportunities, including predictable communication flows for network policies and intrusion detection.
  • Lessons learned: IIoT security mechanisms must support resource-constrained devices and time-critical industrial processes with negligible process-latency impact.The authors identify security protocols with negligible latency impact as a future research need.
  • Lessons learned: Legacy devices that cannot be patched require network-layer access policies and monitoring as complementary security mechanisms.Promising directions include IIoT network traffic policies and process-aware IDSs.
  • Lessons learned: Insider threats require employee awareness and regular training, with future work encouraged on methods such as gamification.The paper states that IIoT and Industry 4.0 remain in their infancy and require new approaches beyond traditional measures.
  • Emerging technologies: Distributed ledgers and blockchains may provide immutable, decentralized accountability, while smart contracts may distributively enforce industrial access control.Security must balance new-device technologies with existing long-lived deployments and legacy hardware.
Loading 2111.11714v1…