Collective intelligence supported by security aware nodes
Project Status: running
Start Date: May 2023
End Date: May 2026
Budget (total): 8482.6 K€
Effort: 74.6 PY
Project-ID: C2022/1-3
Name: Alexey Kirichenko
Company: University of Jyväskylä
Country: Finland
E-mail: alexey.l.kirichenko@jyu.fi
University of Jyväskylä, Finland
Geodata ZT GmbH, Austria
Mattersoft Ltd., Finland
Mint Security Ltd., Finland
Netox, Finland
ScopeSensor Ltd., Finland
Bittium Wireless Ltd., Finland
Bittium Biosignals Ltd., Finland
Nodeon Finland Oy, Finland
Wirepas Oy, Finland
Councilbox, Spain
Affärsverken Karlskrona, Sweden
Arctos Labs AB, Sweden
Clavister AB, Sweden
Blekinge Tekniska Högskolan, Sweden
Blue Science Park, Sweden
Savantic, Sweden
Techinova AB, Sweden
Abstract
The proliferation of Internet of Things (IoT) with its smart devices has fundamentally changed how different environments, such as homes, offices, factories, smart buildings, and smart grids, are used and operated. However, as stated in [1], security is a major concern for IoT networks and environments, where the risks of physical device tampering, injection of malicious devices, and unpatched vulnerabilities are higher than in traditional networks. This is nicely captured in the Hyppönen’s law [2]: “If it’s smart, it’s vulnerable.” Following “when everything is connected, everything must be protected” [2], CISSAN proposes and implements algorithms for mitigating IoT security threats through collective decision-making and with a reduced impact on the limited resources of IoT devices. These algorithms are based on research and innovation in optimizing the distribution of security capabilities and aggregating the intelligence in IoT network nodes. Three industrial use cases, which nowadays heavily rely on the use of IoT, inform the project developments and are used for validating and demonstrating the project results: (i) public transportation; (ii) smart energy grids; (iii) mining and tunnelling operations.
CISSAN algorithms for distributed security monitoring, attack detection and response in IoT networks combine machine learning-based methods, more traditional AI techniques (e.g., decision-making based on formal knowledge representation and expert systems, fuzzy logic-based approaches, or genetic algorithms), and attack-specific rules. Also, methods and tools are developed for verifying the quality of data sets used in the project for building machine learning models and supporting other data-driven technologies. We work to propose network context-aware algorithms for distributing security functions and tasks among IoT devices, edge devices, and possibly cloud backends to achieve a suitable balance between the network resilience and the resource utilization. The technical efforts in CISSAN are accompanied by defining and investigating potential business models around the project results and their business impact analysis. We take into account regulatory and compliance considerations, including the ENISA’s work on certification schemes.
[1] European Cyber Security Organisation (ECSO) Technical Paper on Internet of Things (IoT). Available online: https://ecsorg.eu/?publications=technicalpaper-on-internet-of-things-iot (accessed on 15 January 2024).
[2] Hypponen’s Law: If it’s smart, it’s vulnerable. Available online: https://blog.f-secure.com/hypponens-law-smart-vulnerable/ (accessed on 15 January 2024).