540
B. Farahani et al.
11.6.12 Security and Privacy
IoT e-Health devices, like all networked devices, will present some level of
the potential risk to the security and privacy of end users, through the use of
unauthorized authentications. This is an especially significant concern in the area of
healthcare, where personal safety could be put at risk. In fact, the entire life cycle of
IoT e-Health is built around privacy and security, from specification generation and
all the way to implementation and deployment. Even so, a holistic multi-layered set
of strategies will be necessary in order to overcome the complex security challenges
of engineering an IoT healthcare ecosystem. This approach can be described as
follows [1, 9]:
1. Device layer: Connected devices such as sensors, medical devices, gateways,
fog nodes, and mobile devices, when are involved in capturing, aggregating,
processing and transferring medical data to the cloud. Widespread forms of
attacks in the device layer include tag cloning, spoofing, RF jamming, cloud
polling, and direct connection. In a cloud polling attack, network traffic is
redirected in order to inject commands directly to a device, through the use of
man-in-the-middle (MITM) attacks as well as changes to domain name system
(DNS) configuration. The most effective defense against this attack is an ongoing
policy of evaluation and verification of certifications, at the device level, in order
to ensure that every certificate actually belongs to the e-Health cloud. A direct
connection attack, meanwhile, involves the use of Service Discovery Protocol
like universal plug and play (SSDP/UPnP), or the on-board properties of BLE, to
locate and target IoT devices. This type of attack is best prevented by a policy of
ignoring and blocking unauthenticated requests at the device level, through the
use of robust cryptographic algorithms, along with a key management system.
Other device-layer security measures include identity, authentication, and authorization management, secure booting (i.e., prevent unauthorized applications to
be executed), application sandboxing, whitelisting, fine-grained access control
capability of resources, protection of data during capture, storage, and transit,
traffic filtering feature, fault tolerance, password enforcement policies, secure
pairing protocols, and secure transmission mechanisms. It is also important
to take into account the extremely limited memory, processing capabilities,
power resource, network range, embedded operating systems, and thin embedded
network protocol stacks of many devices while implementing security algorithms
in an IoT Health system [1, 9].
2. Network layer: In this layer, a multitude of diverse network protocols, including
Wi-Fi, BLE, and ZigBee can be leveraged to establish appropriate connections
among sensors. Eavesdropping, Sybil attacks, sinkhole attacks, sleep deprivation
attacks, and Man-in-the-Middle attacks are all common at this level. Thus, the
use of trusted routing mechanisms is crucial, as is the use of message integrity
verification techniques (using hashing mechanisms like MD5 and SHA) and
point-to-point encryption techniques based on cryptographic algorithms. These
algorithms fall broadly into two groups: symmetric algorithms such as AES,
B. Farahani et al.
11.6.12 Security and Privacy
IoT e-Health devices, like all networked devices, will present some level of
the potential risk to the security and privacy of end users, through the use of
unauthorized authentications. This is an especially significant concern in the area of
healthcare, where personal safety could be put at risk. In fact, the entire life cycle of
IoT e-Health is built around privacy and security, from specification generation and
all the way to implementation and deployment. Even so, a holistic multi-layered set
of strategies will be necessary in order to overcome the complex security challenges
of engineering an IoT healthcare ecosystem. This approach can be described as
follows [1, 9]:
1. Device layer: Connected devices such as sensors, medical devices, gateways,
fog nodes, and mobile devices, when are involved in capturing, aggregating,
processing and transferring medical data to the cloud. Widespread forms of
attacks in the device layer include tag cloning, spoofing, RF jamming, cloud
polling, and direct connection. In a cloud polling attack, network traffic is
redirected in order to inject commands directly to a device, through the use of
man-in-the-middle (MITM) attacks as well as changes to domain name system
(DNS) configuration. The most effective defense against this attack is an ongoing
policy of evaluation and verification of certifications, at the device level, in order
to ensure that every certificate actually belongs to the e-Health cloud. A direct
connection attack, meanwhile, involves the use of Service Discovery Protocol
like universal plug and play (SSDP/UPnP), or the on-board properties of BLE, to
locate and target IoT devices. This type of attack is best prevented by a policy of
ignoring and blocking unauthenticated requests at the device level, through the
use of robust cryptographic algorithms, along with a key management system.
Other device-layer security measures include identity, authentication, and authorization management, secure booting (i.e., prevent unauthorized applications to
be executed), application sandboxing, whitelisting, fine-grained access control
capability of resources, protection of data during capture, storage, and transit,
traffic filtering feature, fault tolerance, password enforcement policies, secure
pairing protocols, and secure transmission mechanisms. It is also important
to take into account the extremely limited memory, processing capabilities,
power resource, network range, embedded operating systems, and thin embedded
network protocol stacks of many devices while implementing security algorithms
in an IoT Health system [1, 9].
2. Network layer: In this layer, a multitude of diverse network protocols, including
Wi-Fi, BLE, and ZigBee can be leveraged to establish appropriate connections
among sensors. Eavesdropping, Sybil attacks, sinkhole attacks, sleep deprivation
attacks, and Man-in-the-Middle attacks are all common at this level. Thus, the
use of trusted routing mechanisms is crucial, as is the use of message integrity
verification techniques (using hashing mechanisms like MD5 and SHA) and
point-to-point encryption techniques based on cryptographic algorithms. These
algorithms fall broadly into two groups: symmetric algorithms such as AES,