30
A. M. Said et al.
centralized anomaly-based IDS for smart infrastructures. We chose O-SVM (One
class Support Vector Machines) algorithm for its low energy consuming compared
to other machine learning algorithms for Wireless sensor network (WSN) [20].
As a use case, we are interested in smart hospital infrastructures. Such hospitals have a wide range of resources that are essential to maintain their operations,
patients, employees and the building itself [1,2] safety such as follow:
• Remote care assets: medical equipment for tele-monitoring and tele-diagnosis.
• Networked medical devices: wearable mobile devices (heartbeat bracelet, wireless temperature counters, glucose measuring devices...) or an equipment
installed to collect health service related data.
• Networking equipment: standards equipment providing connectivity between
different equipment (transmission medium, router, gateway...).
• Data: for both clinical and patient data, and staff data, which considered the
most critical asset stored in huge datasets or private clouds.
• Building and facilities: the sensors are distributed in the hospital building
that monitor the patient safety (temperature sensor for patient room and
operation theater, gas sensor are among used sensors).
We target a common IoT architecture that can be considered for smart hospitals. In such architecture, there are mainly three type of components:
Sensing Node: composed of remote care asset, network medical device and
different sensors. These sensors will send different type of data and information
(patient and staff data, medical equipment status...). They are linked to microcontrollers and radio modules to transmit these data to the processing unit [3].
Edge Router: an edge router or border router is a specialized router residing
at the edge or boundary of a network. This node ensures the connectivity of its
network with external networks; a wide area network or the Internet. An edge
router uses an external border gateway protocol, which is used extensively over
the Internet to provide connectivity with remote networks. Instead of providing
communication with an internal network, which the core router already manages,
a gateway may provide communication with different networks and autonomous
systems [4].
Interface Module and Database: this module is the terminal of the network
containing all the collected data from different nodes of the network and analyze those information in order to ensure the safety of patient and improve the
healthcare system.
Figure 1 [5], presents the typical IoT e-health architecture, where sensors are
distributed (medical equipment,room sensors and others) and send data to the
IoT gateway. In one hand, this gives the opportunity to medical supervisor to
control the patient health status. In the other hand, this data will be saved into
databases for more analysis.
The rest of the paper is structured as follows. Section 2 presents the related
work. Section 3 presents the Rank attack scenario. Section 4 presents our proposed approach. Section 5 presents our main results and Sect. 6 concludes the
paper and presents its perspectives.
A. M. Said et al.
centralized anomaly-based IDS for smart infrastructures. We chose O-SVM (One
class Support Vector Machines) algorithm for its low energy consuming compared
to other machine learning algorithms for Wireless sensor network (WSN) [20].
As a use case, we are interested in smart hospital infrastructures. Such hospitals have a wide range of resources that are essential to maintain their operations,
patients, employees and the building itself [1,2] safety such as follow:
• Remote care assets: medical equipment for tele-monitoring and tele-diagnosis.
• Networked medical devices: wearable mobile devices (heartbeat bracelet, wireless temperature counters, glucose measuring devices...) or an equipment
installed to collect health service related data.
• Networking equipment: standards equipment providing connectivity between
different equipment (transmission medium, router, gateway...).
• Data: for both clinical and patient data, and staff data, which considered the
most critical asset stored in huge datasets or private clouds.
• Building and facilities: the sensors are distributed in the hospital building
that monitor the patient safety (temperature sensor for patient room and
operation theater, gas sensor are among used sensors).
We target a common IoT architecture that can be considered for smart hospitals. In such architecture, there are mainly three type of components:
Sensing Node: composed of remote care asset, network medical device and
different sensors. These sensors will send different type of data and information
(patient and staff data, medical equipment status...). They are linked to microcontrollers and radio modules to transmit these data to the processing unit [3].
Edge Router: an edge router or border router is a specialized router residing
at the edge or boundary of a network. This node ensures the connectivity of its
network with external networks; a wide area network or the Internet. An edge
router uses an external border gateway protocol, which is used extensively over
the Internet to provide connectivity with remote networks. Instead of providing
communication with an internal network, which the core router already manages,
a gateway may provide communication with different networks and autonomous
systems [4].
Interface Module and Database: this module is the terminal of the network
containing all the collected data from different nodes of the network and analyze those information in order to ensure the safety of patient and improve the
healthcare system.
Figure 1 [5], presents the typical IoT e-health architecture, where sensors are
distributed (medical equipment,room sensors and others) and send data to the
IoT gateway. In one hand, this gives the opportunity to medical supervisor to
control the patient health status. In the other hand, this data will be saved into
databases for more analysis.
The rest of the paper is structured as follows. Section 2 presents the related
work. Section 3 presents the Rank attack scenario. Section 4 presents our proposed approach. Section 5 presents our main results and Sect. 6 concludes the
paper and presents its perspectives.