A SECURE KEY MANAGEMENT TECHNIQUE FOR WIRELESS BODY AREA NETWORKS

In Wireless Body Area Networks (WBAN), the key factors to be considered for transmission of confidential data are security and privacy as it is mostly having applications in emergency medical response systems. The lack of security may lead to loss of data privacy resulting in an adversary to bring in bogus data or altering the legal ones. Hence in this study, a secure key management technique for WBAN is proposed. The proposed architecture consists of a set of WBANs connected to the master server via backend server using authentication channel. Initially, backend server and master server use a shared symmetric key. When a node wants to join a network, it forwards a request message protected by the Message Authentication Code (MAC) to the master server via the backend server. The master server verifies the MAC and generates message key and master key for the node and sends it to backend server. The backend server encrypts the message key with the master key and sends it to the node that initiates the joining process. After all nodes receive key information from the master server, the Base Server (BS) schedules a re-keying period to refresh the master key. By simulation results, it is shown that the proposed technique is more authenticated. The proposed approach offers data confidentiality and integrity in WBANs.


Wireless Body Area Network (WBAN)
established the concept of WBAN which is otherwise defined as Wireless Personal Area Network (WPAN) (Li et al., 2010). WBAN is the network that permits the combination of smart, small scale, minimum power, aggressive/discreet sensor nodes which monitors body activities and neighboring environment. Every intelligent node in the network has potential to forward the information to the base station after processing to obtain the diagnosis and prescription . The application of WBANs is concerned with medical field and it also upholds consumer electronics applications concurrently (Khan et al., 2009).
The features of WBAN are listed below (Li et al., 2010): • It is a miniature wireless network for communicating within 3 m gap • The speed at which the data is transmitted varies from 10 Kbps to 10 Mbps • The star topology is the fundamental arrangement considered in WBANs and BAN Nodes (BNs) communicate with BAN Network Controller (BNC) alone • BNs possess restricted power, calculation and communication capabilities • Energy efficient security mechanism is required with reduced overhead and also the requirement such as data integrity, authentication and encryption should be fulfilled • The network surrounds the body closely for implanting its communication system • BAN mainly detects, collects and transmits the biomedical information The key issues to be accounted while designing WBAN are Power limitation and short RF transmission range, Mobility, Minimum and time-dependant quality of wireless links and Network size (Nabi et al., 2010).

Security Risks in WBANs
The susceptible nature of wireless channels results in a wide range of security threats distracting the WBAN's Science Publications JCS progress (Saleem et al., 2011). Depending on the network layers, the attacks over WBAN are categorized into following classes: • Physical layer attacks • Data link layer attacks • Network layer attacks • Transport layer attacks

Fundamental Security Requirements in a WBAN
This section presents the fundamental security requirements of WBANs , such as:

Proposed Work
To avoid compromised attacks on preloaded keys and reduce the overhead, a new secure key management mechanism for WBAN is proposed in this study. Tan et al. (2009) have developed a lightweight Identity-Based Encryption (IBELite) suitable for sensors in a BSN. Their protocol balances security and privacy with accessibility. However the proposed system can release n secret keys. When we try to release (n+1) th secret key then the master key is vulnerable to compromise.

Related Work
Venkatasubramanian and Gupta (2010) have presented a novel scheme for securing inter-sensor communication in BSNs called Physiological Value based Security (PVS). Their scheme distributes the key along with the message by hiding it using physiological values.
The proposed security method, however, can be used only when the entire sensor measures the same IPI signals as physiological values which are very difficult. It has more effects of topographic specificity, which has the potential to eliminate key distribution completely. Mohanavalli and Anand (2011) have presented an architecture which continuously monitors the patients at home using BSN. Their scheme facilitates senior citizens, patients with chronic illness and patients who need to be remotely monitored in the comfort of their homes, but this work does not consider end to end security as they use basic authentication techniques. Sharma and Bansal (2011) have examined impersonate attacks in WBAN. They proposed this technique to guarantee whether wireless sensors in a WBAN are transmitting and receiving data from authorized or authentic sources. They utilized digital certificates to overcome Impersonate Attacks. Raazi and Lee (2009) have proposed a distributed key management scheme for WBAN. The keys they used initially for personal server and sensor nodes are preloaded keys, so they can be compromised by adversaries. If initial keys are compromised, then further security operations are useless.

Overview
In this study, a new secure key management mechanism for WBAN is proposed. The proposed architecture consists of a set of WBANs connected to the backend server. The backend server relay the biometric information measured by the sensor node to the master server through the internet. All sensor nodes will discover the master server by using a node id. The master server will then generate a unique secret key for each node. When a node wants to join a network, it forwards a request message protected by the MAC to the master server via the backend server. The master server verifies the MAC and generates a message key and a master key for the node and sends it to backend server. The backend server encrypts message key with master key and sends it to the sensor node that initiates the joining process. After all nodes receive key information from master server, the BS schedules re-keying period to refresh the master key.

Proposed Architecture
Our proposed architecture consists of a set of WBANs, Backend Server (BS) and Master Server (MS).
In Fig. 1 a WBAN network consists of few sensor devices deployed on a human body which are connected to a Backend Server (BS). The BSs of various WBANs are connected to a Master Server (MS).

Backend Server (BS)
The Backend Server (BS) uses three types of keys which are described as follows: Message Key (K msg ) = It is used for communicating with backend server and other nodes Master Key (K mas ) = It is used for refreshing the message key by scheduling the re-keying intervals Secret key (K sec ) = It is shared with the master server and this key is unique for each node

Private Authentication Channel Creation
The techniques described below are used to create a secure out-of-band channel.

Manual Authentication Technique
This technique enables wireless devices and authenticates them through wireless channel for assisting manual data transmission among devices. The manual data transfer involves the operator's actions such as replication of the data output from one device to the other device, comparison of output of two devices, entering similar data into both the devices. This process does not necessitate the user to enter long strings of digits. In general, the user has to enter (or compare) approximately 32 binary digits (Gehrmann et al., 2004).

Distance Bounding Protocols
This protocol merges physical and cryptographic properties and allows the host to estimate the upperbound distance which demands to be inside the transmission range. The estimation of upper bound distance considers the metrics such as Received Signal Strength (RSS), Angle of Arrival (AoA) or Time of Flight (ToF). For securing the distance bounding protocols, the time of flight is taken into consideration (Singelee and Preneel, 2007).

Symmetric Key Generation
For any message transmitted in the network, encryption and authentication are required. Initially, sensor nodes SN i share a symmetric key with MS. The symmetric key i SN sy K for a sensor node SN i is generated as follows Equations 1-4: where, λ = pseudo random function. K sec (SN i ) = secret key of the sensor node After successful authentication, the Master Server (MS) supplies a unique master key (K sec ) to each SN i . K sec holds two sub-keys such as the encryption key (k e ) and the Message Authentication Code (MAC) key (k mac ): When SN i wants to transmit the data to MS, the data is encrypted by the k e and signed by the MAC key K mac before transmission. The format is as follows: where, D is the data, Ts is the timestamp during data transmission and mac (K, d) represents the computation of the message authentication code of message d with key K.
When MS receives the data from any node, a verifies the data and then decrypts it. The secured communication is ensured among the node and MS by the encryption key and MAC key (Wang et al., 2007).

Initialization
Each Sensor Node (SN) is initialized with security by MS prior to joining the WBAN. In this phase, MS and BS use a shared symmetric key. This is performed with the help of private and authenticate out-of band channel. The creation of private and authenticate channel is based on the physical characteristics of the sensor nodes.
Using this technique, the data can be sent through the channel confidentially such that data integrity and authenticity are also protected. The backend server performs like gateway and it is a private authentic channel which provides secure communication between WBAN nodes and master server Initialize the sensor node SN i .
The steps involved in the transfer of the data through the secure out-of-band channel are as follows.
SNi sends its ID to the Master Server (MS): This can be performed in an explicit manner. However the ID of SNi can be known implicitly due to certain properties of the out-of-band channel.
MS generates a random secret key K sec and forwards it to SN i : Both SN i and MS stores K sec in their memory. It is to be noted that each SN i is assigned with a unique secret key K sec . Also it is assigned with a unique Counter (C) with initial value as 0 (CTR→0) and stored in the sensor's buffer. The counter values helps in preventing replay attacks and guarantees consistency. Each time the counter is accessed, the value is incremented by 1.

Node Joining
The new nodes are added to the network during the following circumstances.
• Deployment of new nodes for monitoring certain biometrics • Malfunctioning of a node device The Node joining process is described using the following algorithm. Algorithm 1. SN i forwards a Join Request (JREQ) to the BS:

Re-Keying
When SN i receives key information from the MS, BS schedules re-keying period to refresh the master key. The BS broadcasts the re-keying period along with the new updated K msg , encrypted by K mas to the SN i : Proposed Algorithm: Step 1: The proposed architecture is constructed that includes a set of WBANs, Backend Server (BS) and Master Server (MS). A WBAN network consists of few sensor devices deployed on a human body which are connected to a backend server (BS) Step 2: The backend server and master server agree on a shared symmetric key. All the sensor nodes will discover the master server by using a node id, now master server will generate the unique secret key k sec for each node Step 3: When a node wants to join a network, it will send a request REQ to the backend server. The REQ is protected by Message Authentication Code (MAC), generated by the unique secret key k sec by the joining node Step 4: Upon receiving the REQ message, the backend server forwards the REQ to the master server Step 5: The master server verifies the MAC and generates initial message key k msg and master key k mas for the node and forwards them to backend server Step 6: The backend server encrypts k msg with k mas and sends it to the sensor nodes Step 7: After all nodes receive key information from master server, the BS schedules rekeying period to refresh the master key

Performance Metrics
The BARI (Raazi and Lee, 2009) scheme is considered for performance comparison with SKM. The performance is evaluated based on the following metrics: Average end-to-end delay, Average Packet Delivery ratio and packet drop.

Varying the Keys
Initially the number of keys is varied from 50-250. Figure 2 shows that the average end-to-end delay of our proposed SKM protocol is less than the existing BARI scheme. Figure 3 gives the delivery ratio for both the schemes and it is high for the proposed SKM protocol than the existing BARI scheme. The result of Fig. 4 shows that the packet drop of our proposed SKM protocol is less than the existing BARI scheme. From  Fig. 5, we can see that the overhead of our proposed SKM is lower than the existing BARI scheme.
The average end-to-end delay is represented in Fig. 6. It shows that proposed SKM technique has less delay than the existing BARI scheme. Figure 7 and 8 show the result of packet delivery ratio and packet drop, respectively. We can observe that the delivery ratio is high and packet drop is less for SKM technique when compared to BARI scheme. From Fig. 9, we can see that the overhead of our proposed SKM is lower than the existing BARI scheme.

CONCLUSION
A secure key management technique for WBAN is proposed in this study. The proposed architecture contains a set of WBANs connected to the backend server. The backend server relay the biometric information measured by the sensor node to the master server through the internet. The proposed technique uses shared symmetric key between the backend and master servers in order to ensure authentication of the backend server. The proposed technique is simulated in NS2 and compared with the BARI scheme. The performance is evaluated based on Average end-to-end delay, average packet delivery ratio, packet drop and overhead. Simulation results are in favor of the proposed technique which show improved packet delivery ratio with reduced delay and overhead.