A Lightweight and Robust User Authentication Protocol with User Anonymity for IoT-Based Healthcare

Chien-Ming Chen; Shuangshuang Liu; Shehzad Chaudhry; Yeh-Cheng Chen; Muhammad khan

doi:10.32604/cmes.2022.018749

Chien-Ming Chen1,*, Shuangshuang Liu1, Shehzad Ashraf Chaudhry2, Yeh-Cheng Chen3 and Muhammad Asghar khan4

1College of Computer Science and Engineering, Shandong University of Science and Technology, Qingdao, 266590, China
2Department of Computer Engineering, Istanbul Gelisim University, Istanbul, 34310, Turkey
3Department of Computer Science, University of California, Davis, CA, 95616, USA
4Department of Electrical Engineering, Hamdard University, Islamabad, 44000, Pakistan
*Corresponding Author: Chien-Ming Chen. Email: chienmingchen@ieee.org
Received: 14 August 2021; Accepted: 09 October 2021

In the traditional Internet, most of the information exchange and communication took place between computers, where computers operations were manual operations; the traditional Internet realizes the information exchange and communication between people in a certain sense. Now, however, we have forayed into the era of the Internet of Things (IoT) [1,2]. The applicability of the new system goes beyond realizing the mutual exchange of information and communication between people, between people and objects, and between objects. The IoT has a wide range of uses, including intelligent transportation, intelligent fire protection, intelligent home, intelligent power grid, intelligent medical, and other aspects. In short, it facilitates the use of the latest IT technology in all walks of life. Specifically, IoT technology embeds sensors into the power grid, buildings, and other objects [3 –6]. The construction industry is using IoT technology ubiquitously. Architecture is the foundation of a city; the progress of technology promotes the intelligent development of architecture, and intelligent architecture is rapidly gaining people’s attention. The current smart building methods incorporate power lighting and fire monitoring. Sensors are installed on equipment for sensing, transmission, and remote monitoring, which not only saves considerable time but also energy. Among the many applications of the IoT, smart medicine is one of the most promising applications for the future.

The emergence of IoT technology promotes the further development of medical information technology. IoT technology has great potential in the field of medicine and health [7–9]. It can better realize diagnoses and facilitate intelligent management of things. Furthermore, it realizes digital processing and sharing of resource information, equipment information, drug information, and personnel information. The use of intelligent medicine is prominent in two fields: digital hospitals and medical wearables. The digital hospital includes a hospital information system, medical image storage system, transmission system, and doctor workstation. Their function is to realize the collection, storage, processing, and transmission of patient information. Digital hospitals enable zero-distance contact with patients. Doctors can conduct long-distance consultation, intelligent medical support resource sharing, and cross-regional optimal allocation. In addition, digital medicine can also monitor the vital signs of patients by deploying sensor nodes, which will automatically send an alarm in case of emergency, which reduces the nursing cost of seriously ill patients. The digital hospital also includes a clinical decision-making system, implying that doctors can analyze patients’ symptoms while helping formulate the best and effective treatment plan. In addition, digital medicine provides a remote visitation system. When visitors visit patients, they directly do so through the remote visitation system, which can effectively avoid the direct contact between patients and visitors, eliminate the spread of disease, and shorten the recovery process of patients.

Medical wearable technology [10–12] is the deployment of sensor nodes around the patient, through the sensor nodes collecting information and parameters of the user’s patient and the surrounding environment, sending it through the network to the cloud, and then processing to the user. The digital hospital is an improvement over the traditional hospital; it realizes the digital equipment’s access to electronic medical records and the management of equipment. However, with the introduction of the medical system of the IoT, introducing sensor nodes around the patients to collect information and then transmitting it to remote medical staff is made possible, ensuring the safety of the medical staff. However, the introduction of IoT is bound to involve the transmission of information on the network channel. Due to the universality and openness of the transmission channel, privacy and security of transmitted information have become the main concern of the IoT medical systems.

Fig. 1 shows the architecture of communication between three entities in the IoT-based healthcare environment: the doctor (user), IoT devices, and a gateway. All the IoT devices around the patients collect real-time patients’ information and then transmit them to a gateway. An authenticated doctor can access the gateway to obtain effective information from those IoT devices. This means that a gateway can authenticate the identity of doctors. In other words, a gateway is a medium for doctors to communicate with sensors.

In 2012, Chen et al. [13] proposed an efficient and secure dynamic identity authentication protocol for telemedicine information systems, which dynamically authenticates the user’s identity to achieve user anonymity. However, Cao et al. [14] found that the protocol can track users through offline identity guessing attacks. When the user loses possession of a smart card, there is no guarantee of security as Chen’s protocol is also vulnerable to offline password guessing attacks. Therefore, Cao et al. [14] proposed an improved password authentication protocol based on the smart card. In 2015, He et al. [15] proposed a two-factor authentication scheme for wireless medical sensors, which allows medical personnel to access patient information using wireless sensor medical devices. In 2016, Li et al. [16] proposed a network-based electronic medical authentication scheme, which also uses the user’s password and smart card for two-factor authentication. He et al. [17] proposed an authentication protocol that is more suitable for the configuration of telemedicine information systems with low power consumption mobile devices. Wei et al. [18] found that this protocol cannot effectively resist password attacks; they proposed an improved authentication protocol for telemedicine information systems and proved that the protocol meets the security requirements of two-factor authentication. In 2018, Wu et al. [19] proposed a lightweight two-factor medical authentication scheme, and they claimed that their protocol had perfect security; however, after analysis, it was found that their protocol could not effectively resist perfect forward security. Therefore, based on the two-factor authentication protocol, Wazid et al. [20] proposed a three-factor network authentication key scheme, which introduced biological information based on the previous authentication password and smart card. The map area of biological information is mainly completed by a biological extractor. In 2019, Sharma et al. [21] proposed a lightweight user authentication protocol, but Canetti et al. [22] found that their protocol could not effectively resist privilege insider attacks. Recently, Masud et al. [23] proposed a protocol for the security of the IoT medical system. The paper mentioned that their protocol is a lightweight anonymous user authentication protocol. The protocol only uses hash primitives to encrypt the information, which reduces the burden of the processor while resisting replaying attacks, man-in-the-middle attacks, anonymity, and untraceability. However, we find that the protocol mentioned in this paper cannot effectively resist internal privilege attacks, sensor node capture attacks, or stolen verification attacks, and it cannot provide perfect forward security.

In this paper, we first demonstrate that Masud et al.’s protocol [23] is insecure against various kinds of attacks. We then propose a lightweight and robust user authentication protocol for IoT-based healthcare with user anonymity. In our design, we only use a single hash function and successive XOR operations; thus, the proposed protocol retains better performance. Additionally, the proposed protocol has perfect forward security and can effectively resist internal privilege, stolen verification, and sensor node capture attacks. In addition, we delete some redundant parameters in Masud et al.’s protocol [23] to make the entire protocol more concise. Furthermore, we compare the proposed protocol with other related protocols in terms of communication and computation cost. The results show that our design has better performance. Also, we use the real-or-random (ROR) model [24] to further prove that the proposed protocol is indeed secure.

The remainder of this paper is organized as follows. In Section 2, we briefly review Masud et al.’s protocol [23] Section 3 demonstrates that Masud et al.’s protocol [23] is vulnerable to privilege internal attacks, stolen verification attacks, and sensor node capture attacks. The proposed protocol is described in Section 4. Section 5 and Section 6 provide security and performance analyses and comparisons. Finally, Section 7 concludes the paper.

In this section, we briefly describe the protocol [23], which consists of three phases: user registration phase, sensor node registration phase, and login and mutual authentication phase. In the first two phases, user and sensor registration is conducted through the gateway.

1. The user first selects an DID and password PWD, and then generates a registration request Rreq. Then, the user transmits the DID, PWD, and Rreq to the gateway through the secure channel. After the gateway receives the registration request from the user, it generates a random gateway private key RSG1, calculates

and stores the parameter a, RSG1 and DID in memory. Finally, the gateway returns the calculated parameter a to the user through the secure channel.

2. After receiving the parameter from the gateway, the user first calculates the value of the random gateway key RSG1 according to the parameter a, and then calculates the value of the pseudo-identity DID according to the random private key of the gateway. Secondly, the user encapsulates their password, pseudo-identity, and gateway random private key in parameter B. Finally, the values of user parameters RSG1, DID and B are stored in their own memory. This completes the user’s entire registration process.

1. Firstly, the sensor selects its own identity SID, generates a random sensor private key RSN1, and then transmits the generated parameter SID and RSN1 to the gateway through the secure channel.

2. After receiving the parameters SID, and RSN1, the gateway first generates a random gateway private key RSG2, encapsulates the sensor’s identity, random gateway private key, and random sensor private key in the parameter C through XoR operation, and then calculates

Finally, the gateway stores the values of sensor identity, random sensor private key, random gateway private key, and sensor pseudo-identity in the memory.

This is to test whether the value of Q is equal to B stored in the user memory. If these values are equal, the user generates a temporary random number ND1 and then calculates

Finally, the user transmits parameters ND1∗, DTID, K, and STID to the gateway via a common channel.

2. After receiving the parameter transmitted by the user, the gateway calculates

This is done to verify whether it is equal to the parameter value of K. If it is equal, the gateway generates a temporary random number NG1 and then calculates

Finally, the gateway transmits the parameters GW1, GW2, DTID, SKs, and GW3 to the sensor through the secure channel.

3. The sensor receives the parameters GW1, GW2, DTID, SKs, and GW3 from the gateway and calculates

and then verifies NG1. After verification, it calculates SN1 to verify whether SN1 is equal to the GW2. If it passes verification, the gateway will obtain the session key

Next, the sensor stores the values of RSG2, RSG3, and STIDnew. Finally, the sensor sends the values of SN2, SN3, and SN4 to the gateway through the secure channel.

This verifies whether GW4 is equal to the received value of SN3. If yes, the gateway calculates the values of RSN2 and STIDnew, and then stores the values of RSN2, RSG3, and STIDnew in memory. Next, the gateway generates a random number NG2 and calculates the value of u, SKu, n, and GW5 and then updates the user pseudo-identity. Finally, it stores the values of RSG4 and DTIDnew and transmits the values of parameters u, SKu, n, and GW5 to the user through the common channel.

5. The user obtains the value of NG2 by XOR of the received u and DID and then verifies the NG2. Then, the value of SK and O is computed. Next, the calculated value of O is compared with the value of n. If it is equal, the user continues to calculate the value of RSG4 and DTIDnew. Finally, the user stores the values of RSG4 and DTIDnew in the memory. At this point, the entire login authentication process is complete.

In this section, we first introduce the attack model used in this paper and then analyze Masud’s protocol [23] according to the attack model. The protocol cannot effectively resist privileged insider, sensor node capture, and stolen verification attacks, and there are loopholes in the perfect forward secrecy.

The attack model briefly describes the capabilities of A, which has been described and discussed in [25,26] earlier. The details are as follows:

1. According to the “Dolev-Yao threat (DY) model” [27] proposed before, A can intercept and monitor information through the public channel. In addition, the attacker can modify the transmitted information. In other words, the session messages transmitted between the participants in the protocol through the common channel can be obtained and operated by A. Moreover, A can act as an insider to obtain the information stored in the gateway during the registration phase.

2. Once the sensor is lost and acquired by A, A can use power analysis [28,29] to operate the sensor. The sensitive information stored in the sensor can easily be obtained by A. In this case, if the attacker has additional capabilities, it is easy to carry out sensor simulation and sensor node capture attacks [30].

3. In most user sensor authentication protocols, users often need to store some parameters in the registration phase for use in the login authentication phase. Usually, this information is stored in the user’s smart card or memory. However, the user’s smart card is often easy to lose. Once the smart card is obtained by A, the attacker can use some parameter information stored in the smart card and combine it with some other parameters to carry out a series of attack operations.

A good protocol must comprise the perfect forward secrecy feature [31,32], which ensures that master key leakage will not lead to session key leakage. Forward secrecy can protect past communication from the threat of key exposure in the future. Even in the case of master key leaks, the historical communication still has good security. However, in Masud’s protocol, we found that if A obtains the value of the sensor’s key RSN1, it can conveniently obtain the session key between the gateway and sensor. The specific process is as follows:

2. Second, A intercepts the parameters STID and GW1 through the common channel and then calculates

A can obtain the parameter SKs as it transmits from the gateway to the gateway through the public channel. NG1 can also be calculated through the second step while securing RSN1 in the first step. Therefore, once the sensor key is exposed, the session key is obtained. However, there are some security vulnerabilities in the protocol.

Privileged insider attack refers to a process in which A or the user information administrator obtains some of the user’s basic information and then uses this information to carry out some basic operations, so as to obtain the user key between the medical staff and the sensor node [33].

1. First, A can disguise as a privileged insider. In the process of user registration with the gateway, A can easily obtain the user’s registration information DID and PWD stored in the memory.

2. Second, A intercepts the message u transmitted by the gateway to the user through the common channel and calculates

In the second step, the NG2 is calculated. SKU is transmitted to the user through the common channel in the authentication phase, which can also be obtained by A. The PWD is obtained by A as an insider. Therefore, A can obtain the session key between the user and the gateway. To sum up, Masud’s protocol cannot effectively resist a privilege insider attack.

A stolen verification attack implies that A can decode the value of the session key on the premise of acquiring the information stored in the gateway memory [34]. Masud’s protocol cannot effectively resist the stolen verification attack; the specific attack process is as follows:

1. First, A obtains the parameter RSG1 stored in the gateway memory during user registration, intercepts the parameter DTID sent by the user to the gateway through the public channel during authentication, and then calculates

2. A obtains the parameter RSG4 in the gateway memory during authentication and intercepts the parameter GW5 transmitted by the common channel and then calculates

3. A has calculated the values of NG2 and PWD, and the parameter SKU transmitted through the secure channel. The session key

between the user and the gateway. Therefore, this protocol cannot effectively resist the stolen verification attack.

The sensor node capture attack refers to the process in which the session key is leaked after A obtains the sensors [35]. Through our analysis, we found that Masud’s protocol cannot resist sensor node capture attacks.

1. During the registration of the sensor with the gateway, the sensor stores the identity STID, RSG2∗ and key RSN1 in its own memory. However, sensors are likely to be acquired by A.

2. In the mutual authentication stage of gateway and sensor, the user transmits the parameter STID to the gateway through the common channel, and the gateway transmits the parameter GW1 to the sensor through the common channel. Then, A calculates

Once A obtains the sensor, A can obtain the session keys of both parties through a series of operations.

We have analyzed Masud’s protocol and listed the detailed attack process. A secure protocol must be able to resist some common attacks. We have improved Masud’s protocol, and the improved protocol can successfully repair the aforementioned security vulnerabilities. In addition, we deleted some redundant symbols in the original protocol to make the entire protocol more concise. Our protocol consists of four parts: pre-deployment phase, user registration phase, sensor registration phase, and login authentication phase.

For the pre-deployment phase of users and sensors, the gateway first generates a key-value RG and then sends the key value to the users and sensors through the secure channel in advance.

Fig. 2 illustrates the user registration phase. The detailed steps are as follows:

1. First, the user selects id ID and password PW, and a random number N1, calculates

and generates a request Rreq for registration. Finally, the user transmits the information of RID and RPW to the gateway through the secure channel.

2. After receiving the registration request from the user, the gateway calculates

Finally, A1, A2, and DTID are stored in the gateway, and A2 is transmitted to users through a secure channel.

Fig. 3 illustrates the sensor registration phase. The detailed steps are as follows:

1. First, users select an identity SID for themselves and generate a random key value S1 to calculate

and then the user sends SID′, S1 to the gateway through the secure channel for registration.

2. After receiving the message from the sensor, the gateway generates a random key value G2 and encrypts the key value to obtain

and encrypts the identity of the sensor to obtain the pseudo-identity of the sensor

It next stores the parameters O, B, STID in the gateway memory and then sends the parameter O to the sensor through the secure channel.

3. After the sensor receives the message, it first extracts the value of the gateway’s key G2,

Finally, the sensor stores the parameter value S1 and STID in the sensor memory.

This section introduces the login and mutual authentication process between the user and the sensor through the gateway in detail as in Fig. 4. The following is the detailed description of login and authentication.

1. Before logging in, the user first enters the account id ID and password PW, used in registration. Then, the following is calculated:

Subsequently, D1 is verified to check whether it is equal to the D value previously stored in the user’s memory. If it is, it implies that it is a login operation by a legal user. After successful login, the user calculates

Finally, the user sends the parameter K and SID′ to the gateway through the common channel.

2. When the gateway receives the parameters from a legitimate user, it needs to determine whether the message sent has been tampered with by A, so it calculates

to compare K′ with K, and equality implies that it passes verification. Then, the gateway continues to calculate

SKs is the operation in which the gateway distributes the key to the sensor. Finally, the gateway sends the parameter Y1, Y2, STID, SKs to the sensor through the common channel.

3. The sensor receives the message from the gateway, and first calculates the temporary key value G1 of the gateway according to the values of STID and Y1.

and compares the S2 value sent by the gateway with Y2′. If it is equal, then A has not tampered the parameters sent by the gateway. Next, the sensor calculates the session key

according to the parameter SKs sent by the gateway. Finally, the sensor identity is updated, storing the updated sensor parameter value in memory, and the value of parameter Z is sent to the gateway through the common channel.

4. The gateway receives the parameter from the sensor. First, it checks whether A intercepted the value of the parameter, calculates

and compares the value of Z′ with that of Z. If it is equal, it means it passes verification. Next, the gateway obtains the identity value of the updated sensor through the following operation.

Then, it stores the updated sensor identity value in its memory. Next, the gateway allocates the session key and computes

The gateway stores the updated user’s identity parameter in the gateway, and sends the value of parameter u, SKu, n to the user through the common channel.

If the value of G1′ is equal to the value of G1 previously stored in the user memory, thus passing the verification. Next, the user calculates the session key

between the user and gateway through the value of SKu sent by the gateway. Then, before updating the user’s identity, the following is performed

which is compared with the received value of n. The parameter is updated if it is equal.

In this section, we use the ROR model [24] to prove the security of the proposed protocol. In the protocol, we define three entities: user, gateway, and sensor node. For this proof, we assume that Ui, Gj, and Sz are the i-th user, the j-th gateway, and the z-th sensor node, respectively, and the parameter T={Ui,Gj,Sz}. In the initial stage, A can perform the following query operations.

Execute (T): By performing this operation, A can obtain the messages {K}, {Y1,Y2,STID,SKs}, {Z}, and {u,SKu,n} transmitted by U, G, and S through the common channel.

CorruptDevice (T): After executing this query, A can get the information stored in the U, G and S′ memory. In addition, A can also get the long-term key in the protocol and the temporary information generated by the participant.

Hash (string): After entering a fixed-length string, A can get a fixed value after executing the query.

Test (T): In the initial stage, A tosses a coin O with uniform texture to judge whether the obtained session key is correct. If O = 1, the session key obtained is correct. Otherwise, A obtains a string with the same length as the session key.

Theorem: For the ROR model, if A performs some basic query operations, the probability that it can break the proposed protocol T in polynomial time is

In the formula, f represents the length of biological information entered by the user in the registration and login stage, and D′ and b′ represent two constants.

Proof: We defined 6 games GM0 to GM5 in the specific proof process, and everyone has different game rules. In the proof process, Succ(GMi)(η) (i = 0, 1, 2, 3, 4, 5) represents the probability of game success under each rule. The specific proof process is as follows:

GM0: In this game, A does not perform any query operations, so the probability of it breaking protocol T is: AdvAT(η)=|2Pr[SuccAT(η)−1]|.

GM1: GM1 adds the Execute query operation on the basis of Gm0. That is, A can obtain the information M1, M2, M3 and M4 transmitted through the common channel. Then, it obtains the session key SK through the Test operation, but the random key S1 of the S and G cannot be obtained, so the probability of success of GM1 is equal to that of GM0.

GM2: GM2 adds the Send operation on the basis of GM1, that is, A can send information to participants through the public channel. Under Zipf’s law, we can easily obtain

GM3: GM3 adds Hash query on the basis of GM2. A can get specific values through Hash operation. According to the birthday paradox, we get

GM4: In this game, we query the CorruptDevice to obtain the value of long-term key RG and the value of temporary information N1 generated by U to verify whether the protocol has perfect forward security and resists temporary information leakage attacks.

Perfect forward secrecy: A obtains the parameter RG through the CorruptDevice operation, but it cannot obtain the user’s pseudo identity RID, so the values of parameter A2 and the user’s pseudo password RPW cannot be obtained. Therefore, A obtains the long-term key, and RG cannot successfully obtain the session key.

Temporary information leakage attack: A obtains the temporary information N1 generated by U but cannot obtain the user’s identity ID and password PW. Therefore, A cannot obtain the user’s pseudo password RPW. Even if A obtains the temporary information N1, it cannot successfully obtain the session key. Therefore, our probability of getting GM4 is

GM5: Different from the GM4 rule, we query the information stored in the user’s memory through CorruptDevice, and then prove that the proposed protocol can resist offline password guessing attacks. The probability that it can successfully guess the user password is 1/2, but in Zipf’s law, when the number of transmitted bits qsend≤106, the probability that A can successfully guess the user password is greater than 1/2. Therefore, we get

GM6: In GM6, in order to verify that the protocol we proposed can successfully resist user simulation attacks, unlike GM5, A queries through Hash operation. Therefore, the probability of GM6 is

Because the probabilities of GM6 success and failure are equal, Pr[SuccAGM6]=1/2.

According to the above process, we prove that our proposed protocol can effectively resist user simulation, offline password guessing, and temporary information leakage attacks and has perfect forward security.

In this section, we describe how the new protocol can resist several common attacks. The following descriptions further prove the security of our proposed protocol.

In this protocol, we assume that A disguises itself as a privileged insider. Therefore, A can obtain the user’s pseudo-identity RID and pseudo password RPW. However, parameter A2 is obtained after RG encryption. A2=RID⊕RPW⊕G1⊕RG, RG is the long-term key generated by the gateway, which is only transmitted to users and sensors in the pre-deployment phase, so only users, gateways, and sensors know the key value. As a privileged insider, A cannot obtain the long-term key value, and thus cannot obtain A2 and the session key between the user and the gateway.

If A captures the sensor node information STID and S1, then, although A already knows the parameters STID and S1, it must know the long-term key RG of the gateway to obtain the parameter O. The calculation of the parameter RG must be participated by G2 and B. B calculation is obtained by long-term key RG of the gateway, even if the node information of the sensor is captured, it impossible to obtain the public key between the gateway and the sensor. Therefore, our improved protocol can effectively resist the sensor capture attack.

In a stolen verification attack, A obtains the message in the gateway memory, and A can obtain the key of both sides of the session. Suppose A obtains the information A1, A2, DTID and RID in the gateway memory. First, parameters Y1 and STID are transmitted on the common channel so we can obtain the value of parameter G1. According to G1 and DTID, we can obtain RID. However, to calculate RPW, we must know the long-term key RG of a gateway. However, the acquisition of RG must be participated by RPW. Therefore, A cannot obtain the session key between the user and the gateway effectively. Suppose A obtains the information O, B, STID and SID′ in the gateway; then, it can get the parameter G1 according to the obtained information, but the session key is also composed of G2 and S1. The sensor’s temporary key value S1 is generated after encryption by the gateway’s temporary key G2 and long-term key RG, and it is impossible to obtain S1. Therefore, A cannot successfully obtain the public session key between the gateway and the sensor. In conclusion, our new protocol can successfully resist the stolen authentication attack.

Assuming that A has obtained the long-term key RG of the gateway, for the public session key between the user and the gateway, A needs to know the parameter A2 and the user’s pseudo password RPW. However, it does not know the user’s pseudo-identity RID, so it cannot obtain the parameters A2 and RPW, so the session key between the user and the gateway can be effectively protected. Second, for the session key between the sensor and the gateway, even if A obtains the long-term key RG, the communication between the gateway and the sensor still requires S1, G2 and O. G2 needs to be obtained through the pseudo-identity of the sensor, but the pseudo-identity of the sensor cannot be obtained, and O must be obtained by the participation of the sensor’s temporary key value S1; therefore, A cannot obtain the session key between the sensor and the gateway.

In the user registration stage, we perform the XOR operation on the user’s ID and the long-term key RG of the gateway and then encrypt the user’s identity. Subsequently, communication with the gateway occurs through the secure channel. Therefore, it is not easy for A to obtain the identity of legitimate users, so our protocol protects the identity privacy of users.

In the user login phase, the system verifies whether the value of D1 is equal to the value of D stored in the user memory. A guesses the identity of a legitimate user if it can successfully guess the user’s password. However, the user authentication also needs the participation of the random number N1 generated by the user in the registration phase, so A cannot successfully carry out a password guessing attack.

If A obtains the random N1 generated by the user in the registration phase but does not know the user’s ID and password PW, the user’s pseudo password RPW cannot be obtained. However, the session key between the user and the gateway needs the participation of the user’s pseudo password RPW. Therefore, A cannot successfully carry out the temporary information leakage attack.

In this section, we analyze the security and performance of the new protocol. We compare the new protocol with other related protocols, mainly by comparing the running time, communication cost, and the ability to resist common attacks to show that our proposed protocol has an advantage in security and performance.

In this part, we compare with other related agreements. Finally, other protocols cannot resist all common attacks, but our new protocol can resist all attacks. At present, common network attacks include A1: Identity anonymity of user device, A2: Identity anonymity of IoT sensor node, A3: privileged-insider attack, A4: off-line password guessing attack, A5: Perfect forward secrecy, A6: man-in-the-middle attack, A7: IoT sensor node impersonation attack, A8: Sensor node capture attack, A9: Stolen verification attack. The comparison results are presented in Table 2. A “Yes” implies that the protocol can resist the attack, whereas a “No” means that it cannot.

For performance analysis, we use the same conditions to analyze the protocols in different environments. In the analysis process, because XOR and join operations take less time, we only analyze according to the non-collision hash function used in the protocol. The time required for the hash function is 0.00089 Ms. In addition, in the communication process, the number of bits required for the non-collision hash function is 256 bits.

First, we compare the communication cost between the protocol proposed in this paper and the related protocols proposed earlier. Here, we only consider the communication cost of the non-collision hash function. The communication cost of our protocol is 1,792 bits, lower than those of Masud et al. [23] (4,096 bits), Wazid et al. [40] (2,304 bits), Turkanovi et al. [39] (5,120 bits), Farash et al. [38] (5,888), Zhou et al. [37] (6,144 bits), and Challa et al. [36] (3,840 bits). This result can be observed in Fig. 5.

Second, we compare the protocols proposed in this paper with regard to time. Here, we only consider the running time of the non-collision hash function. Table 3 shows the number of hash functions required by the user gateway and sensor nodes during the protocol user registration phase, sensor registration phase, and login authentication phase where H represents the hash function. In Table 4, we compare the proposed protocol with those in other related fields. The results show that the time required for our proposed protocol is 0.00623 ms, and for Masud et al. [23], Wazid et al. [40], Turkanovi et al. [39], Sharma et al. [21], Farash et al. [38], Zhou et al. [37], Challa et al. [36], the times are 0.00712, 0.02848, 0.01513, 0.02047, 0.02848, 0.03204, and 0.01068 ms, respectively. It can be seen more intuitively in Fig. 6 that the running cost of the protocol proposed by us is better than those proposed in other relevant papers.

After comparing our protocol with other related protocols, we can observe that the proposed protocol can effectively resist various attacks, and so we can say that our protocol has perfect security. In addition, our proposed protocol is superior to the existing protocol in terms of communication cost and time running cost. To sum up, the proposed protocol is more suitable for the development of future medical systems and is more convenient and user friendly for future medical staff and patients.

This paper improves Masud’s authentication protocol for the medical system. The improved protocol not only resists the common attacks that the existing protocol was unable to but also removes the redundant symbols in the original protocol, reducing the communication cost. In addition, it retains the lightweight advantage of the original protocol. The improved protocol still adopts a single hash and bit-by-bit XOR operation, which reduces the running time. The protocol is secure against privileged internal attacks, stolen verification attacks, and sensor node capture attacks, thus presenting perfect forward security. This protocol is more suitable for the future medical environment. It preserves the security in the medical system as well as the user privacy, while additionally enhancing the system performance.

Conflicts of Interest: The authors declare that they have no conflicts of interest to report regarding the present study.

References

Xiong, H., Huang, X., Yang, M., Wang, L., Yu, S. (2021). Unbounded and efficient revocable attribute-based encryption with adaptive security for cloud-assisted Internet of Things. IEEE Internet of Things Journal. DOI 10.1109/JIOT.2021.3094323. [CrossRef]
Hou, Y., Xiong, H., Huang, X., & Kumari, S. (2021). Certificate-based parallel key-insulated aggregate signature against fully chosen key attacks for industrial Internet of Things. IEEE Internet of Things Journal, 8(11), 8935-8948. [Google Scholar] [CrossRef]
Wang, K., Chen, C. M., Liang, Z., Hassan, M. M., & Sarné, G. M. (2021). A trusted consensus fusion scheme for decentralized collaborated learning in massive IoT domain. Information Fusion, 72, 100-109. [Google Scholar] [CrossRef]
Wang, P., Chen, C. M., Kumari, S., Shojafar, M., & Tafazolli, R. (2021). HDMA: Hybrid D2D message authentication scheme for 5G-enabled vanets. IEEE Transactions on Intelligent Transportation Systems, 22(8), 5071-5080. [Google Scholar] [CrossRef]
Wang, X., Liu, Y., & Choo, K. K. R. (2020). Fault-tolerant multisubset aggregation scheme for smart grid. IEEE Transactions on Industrial Informatics, 17(6), 4065-4072. [Google Scholar] [CrossRef]
Li, C. T., Lee, C. C., Weng, C. Y., & Chen, C. M. (2018). Towards secure authenticating of cache in the reader for rfid-based IoT systems. Peer-to-Peer Networking and Applications, 11(1), 198-208. [Google Scholar] [CrossRef]
Wu, T. Y., Wang, T., Lee, Y. Q., Zheng, W., & Kumari, S. (2021). Improved authenticated key agreement scheme for fog-driven IoT healthcare system. Security and Communication Networks, 2021, [Google Scholar]
Chen, C. M., Li, C. T., Liu, S., Wu, T. Y., & Pan, J. S. (2017). A provable secure private data delegation scheme for mountaineering events in emergency system. IEEE Access, 5, 3410-3422. [Google Scholar] [CrossRef]
Ayub, M. F., Mahmood, K., Kumari, S., & Sangaiah, A. K. (2021). Lightweight authentication protocol for e-health clouds in IoT-based applications through 5G technology. Digital Communications and Networks, 7(2), 235-244. [Google Scholar] [CrossRef]
Shahbazi, Z., & Byun, Y. C. (2020). Towards a secure thermal-energy aware routing protocol in wireless body area network based on blockchain technology. Sensors, 20(12), 3604. [Google Scholar] [CrossRef]
Benmansour, T., Ahmed, T., Moussaoui, S., & Doukha, Z. (2020). Performance analyses of the IEEE 802.15. 6 wireless body area network with heterogeneous traffic. Journal of Network and Computer Applications, 163, 102651. [Google Scholar] [CrossRef]
Hasan, K., Biswas, K., Ahmed, K., Nafi, N. S., & Islam, M. S. (2019). A comprehensive review of wireless body area network. Journal of Network and Computer Applications, 143, 178-198. [Google Scholar] [CrossRef]
Chen, H. M., Lo, J. W., & Yeh, C. K. (2012). An efficient and secure dynamic ID-based authentication scheme for telecare medical information systems. Journal of Medical Systems, 36(6), 3907-3915. [Google Scholar] [CrossRef]
Cao, T., & Zhai, J. (2013). Improved dynamic ID-based authentication scheme for telecare medical information systems. Journal of Medical Systems, 37(2), 9912. [Google Scholar] [CrossRef]
He, D., Kumar, N., Chen, J., Lee, C. C., & Chilamkurti, N. (2015). Robust anonymous authentication protocol for health-care applications using wireless medical sensor networks. Multimedia Systems, 21(1), 49-60. [Google Scholar] [CrossRef]
Li, X., Niu, J., Karuppiah, M., Kumari, S., & Wu, F. (2016). Secure and efficient two-factor user authentication scheme with user anonymity for network based e-health care applications. Journal of Medical Systems, 40(12), 1-12. [Google Scholar] [CrossRef]
He, D. B., Chen, J. H., & Zhang, R. (2012). A more secure authentication scheme for telecare medicine information systems. Journal of Medical Systems, 36(3), 1989-1995. [Google Scholar] [CrossRef]
Wei, J., Hu, X., & Liu, W. (2012). An improved authentication scheme for telecare medicine information systems. Journal of Medical Systems, 36(6), 3597-3604. [Google Scholar] [CrossRef]
Wu, F., Li, X., Sangaiah, A. K., Xu, L., & Kumari, S. (2018). A lightweight and robust two-factor authentication scheme for personalized healthcare systems using wireless medical sensor networks. Future Generation Computer Systems, 82, 727-737. [Google Scholar] [CrossRef]
Wazid, M., Das, A. K., & Vasilakos, A. V. (2018). Authenticated key management protocol for cloud-assisted body area sensor networks. Journal of Network and Computer Applications, 123, 112-126. [Google Scholar] [CrossRef]
Sharma, G., & Kalra, S. (2019). A lightweight user authentication scheme for cloud-IoT based healthcare services. Iranian Journal of Science and Technology, Transactions of Electrical Engineering, 43(1), 619-636. [Google Scholar] [CrossRef]
Canetti, R., Krawczyk, H. (2002). Universally composable notions of key exchange and secure channels. International Conference on the Theory and Applications of Cryptographic Techniques, Springer, The Netherlands.
Masud, M., Gaba, G. S., Choudhary, K., Hossain, M. S., Alhamid, M. F. et al. (2021). Lightweight and anonymity-preserving user authentication scheme for IoT-based healthcare. IEEE Internet of Things Journal. DOI 10.1109/JIOT.2021.3080461. [CrossRef]
Abdalla, M., Fouque, P. A., Pointcheval, D. (2005). Password-based authenticated key exchange in the three-party setting. International Workshop on Public Key Cryptography, Springer.
Wang, D., He, D., Wang, P., & Chu, C. H. (2014). Anonymous two-factor authentication in distributed systems: Certain goals are beyond attainment. IEEE Transactions on Dependable and Secure Computing, 12(4), 428-442. [Google Scholar] [CrossRef]
Wang, D., & Wang, P. (2016). Two birds with one stone: Two-factor authentication with security beyond conventional bound. IEEE Transactions on Dependable and Secure Computing, 15(4), 708-722. [Google Scholar] [CrossRef]
Dolev, D., & Yao, A. (1983). On the security of public key protocols. IEEE Transactions on Information Theory, 29(2), 198-208. [Google Scholar] [CrossRef]
Kocher, P., Jaffe, J., Jun, B. (1999). Differential power analysis. Annual International Cryptology Conference, Springer. California, USA.
Ren, Y., Wu, L. (2013). Power analysis attacks on wireless sensor nodes using CPU smart card. 22nd Wireless and Optical Communication Conference, IEEE, Chongqing, China.
Far, H. A. N., Bayat, M., Das, A. K., Fotouhi, M., & Pournaghi, S. M. (2021). LAPTAs: Lightweight anonymous privacy-preserving three-factor authentication scheme for WSN-based IIoT. Wireless Networks, 27(2), 1389-1412. [Google Scholar] [CrossRef]
Li, P., Su, J., & Wang, X. (2020). iTLS: Lightweight transport-layer security protocol for IoT with minimal latency and perfect forward secrecy. IEEE Internet of Things Journal, 7(8), 6828-6841. [Google Scholar] [CrossRef]
Xu, S., Liu, X., Ma, M., & Chen, J. (2020). An improved mutual authentication protocol based on perfect forward secrecy for satellite communications. International Journal of Satellite Communications and Networking, 38(1), 62-73. [Google Scholar] [CrossRef]
Ul Haq, I., Wang, J., Zhu, Y., & Maqbool, S. (2020). A survey of authenticated key agreement protocols for multi-server architecture. Journal of Information Security and Applications, 55, 102639. [Google Scholar] [CrossRef]
Chen, C. M., & Ku, W. C. (2002). Stolen-verifier attack on two new strong-password authentication protocols. IEICE Transactions on Communications, 85(11), 2519-2521. [Google Scholar]
Xu, G., Wang, F., Zhang, M., & Peng, J. (2020). Efficient and provably secure anonymous user authentication scheme for patient monitoring using wireless medical sensor networks. IEEE Access, 8, 47282-47294. [Google Scholar] [CrossRef]
Challa, S., Wazid, M., Das, A. K., Kumar, N., & Reddy, A. G. (2017). Secure signature-based authenticated key establishment scheme for future IoT applications. IEEE Access, 5, 3028-3043. [Google Scholar] [CrossRef]
Zhou, L., Li, X., Yeh, K. H., Su, C., & Chiu, W. (2019). Lightweight IoT-based authentication scheme in cloud computing circumstance. Future Generation Computer Systems, 91, 244-251. [Google Scholar] [CrossRef]
Farash, M. S., Turkanovi, M., Kumari, S., & Hölbl, M. (2016). An efficient user authentication and key agreement scheme for heterogeneous wireless sensor network tailored for the Internet of Things environment. Ad Hoc Networks, 36, 152-176. [Google Scholar] [CrossRef]
Turkanović, M., Brumen, B., & Hölbl, M. (2014). A novel user authentication and key agreement scheme for heterogeneous ad hoc wireless sensor networks, based on the Internet of Things notion. Ad Hoc Networks, 20, 96-112. [Google Scholar] [CrossRef]
Wazid, M., Das, A. K., Shetty, S., Rodrigues, J. J. P. C., & Park, Y. (2019). LDAKM-EIoT: Lightweight device authentication and key management mechanism for edge-based IoT deployment. Sensors, 19(24), 5539. [Google Scholar] [CrossRef]