**2.3. Security vulnerabilities**

In most of applications, sensor devices are spread over large areas, what difficult a individual control of network components. Moreover, wireless communication allows an attacker can trigger attacks without having physical access to the device, so according to Shi and Perrig [7] attacks on WSNs can be divided into three main types: (1) Attack of authentication and confidentiality: Consists of attacks change, repetition or modification packages. (2) Availability network Attack: Generally known as DoS attacks or negation of service, this attack involves the application of techniques that make the network unavailable. (3)Attack on integrity: this type of attack the attacker's goal is to inject false data on the network, keeping the network available, but traveling fictitious data. Table 1 described by Wang [8] illustrates the most common types of attacks in WSN considering the network layer in which they operate.


**Table 1.** Possible attacks on a Wireless Sensor Network [8]

At the physical layer can occur the following attacks: jamming and tampering. The attack jamming consists in the interference of radio frequency signal that sensor nodes use to communicate. The tampering attack occurs due to physical vulnerability of sensor nodes spread over large areas, therefore susceptible to capture, breaking the circuit, setting modification or even replacement of a network node by a malicious sensor node [9]. At link layer attacks can be of the collision, when two sensor nodes attempt to transmit while at the same frequency, in this case the packet is discarded and must be retransmitted [10]. The attacker may cause intentional collisions by a malicious sensor node. Repeated collisions can lead to exhaustion of resources, making it unavailable sensor nodes. Also in the link layer unfairness attack is a type of DoS when the adversary causes degradation of real-time applications run on other sensor nodes by intermittent interruption of the transmission of their frames.

Denial of Service (DoS) attacks consist of flooding the receiver with no other requests for communication can be performed during the attack, leaving the involved nodes unavailable for new connections.

In the network layer attacks can occur of type Spoofed Routing Information, where the attacker modifies routing table information. The routes make false packets do not reach the correct destination, or even make the referral to consume more resources than normal [11]. The Selective Forwarding attack is the involvement of a sensor node by an attacker who causes some messages to be routed and other discarded [11]. In the Sinkhole attack the attacker causes a compromised sensor node is seen as most efficient route to the sink of the network, thus the neighboring nodes will always use the attacker to send their data [12][11][10].

4 Will-be-set-by-IN-TECH

• **Military -** To detect the presence of enemies, explosions, presence of hazardous materials

In most of applications, sensor devices are spread over large areas, what difficult a individual control of network components. Moreover, wireless communication allows an attacker can trigger attacks without having physical access to the device, so according to Shi and Perrig [7] attacks on WSNs can be divided into three main types: (1) Attack of authentication and confidentiality: Consists of attacks change, repetition or modification packages. (2) Availability network Attack: Generally known as DoS attacks or negation of service, this attack involves the application of techniques that make the network unavailable. (3)Attack on integrity: this type of attack the attacker's goal is to inject false data on the network, keeping the network available, but traveling fictitious data. Table 1 described by Wang [8] illustrates the most common types of attacks in WSN considering the network layer in which

• **Security -** To provide security in homes, shopping centers, farms, among others.

**Layer Types of Attacks**

Application Layer *Malicious Node*

**Table 1.** Possible attacks on a Wireless Sensor Network [8]

Physical Layer *jamming* ou ataque de interferncia Link Layer *collision, exhaustion, unfairness*

*flood, Ack Flooding,*

Transport Layer *Flooding De-synchronization*

Network Layer *spoofed routing information and selective*

At the physical layer can occur the following attacks: jamming and tampering. The attack jamming consists in the interference of radio frequency signal that sensor nodes use to communicate. The tampering attack occurs due to physical vulnerability of sensor nodes spread over large areas, therefore susceptible to capture, breaking the circuit, setting modification or even replacement of a network node by a malicious sensor node [9]. At link layer attacks can be of the collision, when two sensor nodes attempt to transmit while at the same frequency, in this case the packet is discarded and must be retransmitted [10]. The attacker may cause intentional collisions by a malicious sensor node. Repeated collisions can lead to exhaustion of resources, making it unavailable sensor nodes. Also in the link layer unfairness attack is a type of DoS when the adversary causes degradation of real-time applications run on other sensor nodes by intermittent interruption of the transmission of

Denial of Service (DoS) attacks consist of flooding the receiver with no other requests for communication can be performed during the attack, leaving the involved nodes unavailable

In the network layer attacks can occur of type Spoofed Routing Information, where the attacker modifies routing table information. The routes make false packets do not reach the correct destination, or even make the referral to consume more resources than normal [11]. The Selective Forwarding attack is the involvement of a sensor node by an attacker who causes

*forwarding, sinkhole, sybil, wormhole, Hello*

as poison gas and radiation.

**2.3. Security vulnerabilities**

they operate.

their frames.

for new connections.

The Sybil attack happens when a malicious node takes over a network identity. According to Douceur [13] this attack was originally intended for distributed systems of redundant data storage, but it is also effective against routing algorithms, data aggregation, and resource allocation, among others. The Wormhole attack consists in a low latency link between two sensor nodes of a network through which an attacker generates messages with court order to exhaust the resources of the devices [11]. In the Hello Flood attack the attacker can use a high power transmitter to fool a large number of sensor nodes, making them believe they are close [11].

Subsequently the attacker sends a fake shortest path to base station, and all nodes receiving Hello packets, try to convey through the attacking node. However, these nodes are out of radio range of the malicious node. Some routing information algorithms use state of sensor nodes. The Acknowledgment Spoofing attack consists in spreading false information about the states of neighboring sensor nodes performed by a malicious sensor node in order to prevent packets from reaching their destinations [11].

In the transport layer, Flooding attack consists in the flood of requests to new connections in order to exhaust the resources of memory and prevent the closure of legitimate requirements of provisions. The De-synchronization attack refers to the interruption of an existing connection [10]. In this attack the attacker captures messages forcing the sender to resend them expending energy unnecessarily.

There are also attacks that exploit vulnerabilities in authentication and data confidentiality. The attack consists of setting replication of a malicious node assumes the identity of a network node. This false node can forward packets in corrupt or false routes. If the attacker has physical access to network, it can copy cryptographic keys and use them in false messages. Also the attacker can deploy the malicious node at strategic locations in order to divide the WSN.

Preserving privacy in data transmission in WSN is challenging, since this type of network allows remote access. Moreover, a single adversary can monitor multiple networks simultaneously [14]. Eavesdropping and passive monitoring are the most common and easiest attack to data privacy. In this type of attack the spy monitors the data transfer and can access its contents if no encryption mechanism implemented in the network being monitored. The traffic analysis is usually applied in conjunction with the attack of listening and passive monitoring. It consists of the preliminary analysis of network traffic to identify nodes that are generating data exchange that interest to the attacker. Finally, the camouflage attack, wherein the malicious attacker deploys a node in the network forwards packets to sensor nodes being monitored.

Through this analysis one can see that there is a range of attacks for WSNs in all layers of the TCP / IP protocol stack. Furthermore, it is apparent that a common point in most attacks is the exploitation of low computing capacity of sensor nodes, as are injected false data and routes are always altered in order to occupy the lower transmission capacity of the sensor nodes, or eliminate its reserve energy. Others attacks yet unidentified may occur in WSN, and protect the network from these threats can be a difficult task.

### 6 Will-be-set-by-IN-TECH 222 Wireless Sensor Networks – Technology and Protocols
