@misc{10498/29841,
year = {2020},
month = {7},
url = {http://hdl.handle.net/10498/29841},
abstract = {The Internet of Things (IoT) is growing globally at a fast pace: people now find themselves surrounded by a variety of IoT devices such as smartphones and wearables in their everyday lives. Additionally, smart environments, such as smart healthcare systems, smart industries and smart cities, benefit from sensors and actuators interconnected through the IoT. However, the increase in IoT devices has brought with it the challenge of promptly detecting and combating the cybersecurity attacks and threats that target them, including malware, privacy breaches and denial of service attacks, among others. To tackle this challenge, this paper proposes an intelligent architecture that integrates Complex Event Processing (CEP) technology and the Machine Learning (ML) paradigm in order to detect different types of IoT security attacks in real time. In particular, such an architecture is capable of easily managing event patterns whose conditions depend on values obtained by ML algorithms. Additionally, a model-driven graphical tool for security attack pattern definition and automatic code generation is provided, hiding all the complexity derived from implementation details from domain experts. The proposed architecture has been applied in the case of a healthcare IoT network to validate its ability to detect attacks made by malicious devices. The results obtained demonstrate that this architecture satisfactorily fulfils its objectives.},
abstract = {El Internet de las Cosas (IoT) está creciendo a nivel global a un ritmo acelerado: las personas ahora se encuentran rodeadas de una variedad de dispositivos IoT como smartphones y wearables en su vida cotidiana. Además, los entornos inteligentes, como los sistemas de atención médica inteligentes, las industrias inteligentes y las ciudades inteligentes, se benefician de sensores y actuadores interconectados a través del IoT. Sin embargo, el aumento de los dispositivos IoT ha traído consigo el desafío de detectar y combatir rápidamente los ataques y amenazas de ciberseguridad que los tienen como objetivo, incluyendo malware, violaciones de privacidad y ataques de denegación de servicio, entre otros. Para abordar este desafío, este documento propone una arquitectura inteligente que integra la tecnología de Procesamiento de Eventos Complejos (CEP) y el paradigma de Aprendizaje Automático (ML) con el fin de detectar diferentes tipos de ataques de seguridad en IoT en tiempo real. En particular, dicha arquitectura es capaz de gestionar fácilmente patrones de eventos cuyas condiciones dependen de los valores obtenidos por los algoritmos de ML. Además, se proporciona una herramienta gráfica impulsada por modelos para la definición de patrones de ataque de seguridad y la generación automática de código, ocultando toda la complejidad derivada de los detalles de implementación a los expertos del dominio. La arquitectura propuesta ha sido aplicada en el caso de una red de IoT de atención médica para validar su capacidad para detectar ataques realizados por dispositivos maliciosos. Los resultados obtenidos demuestran que esta arquitectura cumple satisfactoriamente sus objetivos.},
organization = {This work was supported by the Spanish Ministry of Science, Innovation and Universities and the European Union FEDER Funds [grant numbers FPU 17/02007, RTI2018-093608-B-C33, RTI2018- 098156-B-C52 and RED2018-102654-T]. This work was also sup- ported by the JCCM [grant number SB-PLY/17/180501/ 0 0 0353 ]},
publisher = {Elsevier Ltd.},
keywords = {Complex event processing},
keywords = {Machine learning},
keywords = {Software architecture},
keywords = {Intelligent decision making},
keywords = {Internet of Things},
keywords = {Security attack},
title = {Integrating complex event processing and machine learning: An intelligent architecture for detecting IoT security attacks},
doi = {10.1016/j.eswa.2020.113251},
author = {Roldán, José and Boubeta Puig, Juan and Martínez, José Luis and Ortiz Bellot, Guadalupe},
}