A Novel Fog Computing-based Telehealth System With SMS And Homomorphic Encryption

Background Information

Title: A Novel fog computing-based telehealth with SMS and homomorphic encryption

In this section a lot will not be canvased as it will set precedent on what is going to be done on chapter one. Firstly this is where we get reflection of what kind of the problem that the user is trying to achieve though we had small hit provided in the abstract.

In this chapter we intend to cover some of the key thematic areas which include and not limited to the following; problem statement, solution proposal to the problem we intend to solve, Aims and objectives, Research questions, Research justification and Research scope and limitation.

Health being one kind of industry dealing with critical data, real-time and confidential data systems that are in place should be very accurate and reliable at all times. Once data has been conveyed, it should land in the authorized hands. The patient record is so critical it needs to be handled with care. Data protection laws also emphasise the privacy of individual data[10]. It could be very unpleasant that private, such as medical conditions, has been exposed to the public. This demoralises the patient or any other person in essence. Therefore, we should ensure patient data confidentiality, availability, and integrity.

Therefore, when you check the data in question, the security of the data is much more critical in all aspects because of the information in real-time. Though the data is real-time, data require security features to travel across the network. Digital assailants being on the ground and what to impersonate or data session, we must institute a mechanism of ensuring and enhancing the security of the data.

To ensure the security of the data, the research adopts homomorphic encryption standards (HES) to encrypt our data. It will enhance the security of our data sent across the networks. Additionally, we are proposing fog computing, one of the services offered in the cloud-like middleware interface [9]. Data transmitted across the wireless network is now secure due to homomorphic encryption, and fog computing provides real-time patient data delivery and retrieval.

Moreover, fog computing is of more significant advantage as a security enhancement to our medical data. It is a distributed framework in which workload is distributed to different geographically located servers. However, it should be noted that most of the fog servers are located in an environment where we have security threats. Anticipating that our data security is at stake[3][6]. To mitigate such kind of threat, we always encrypt our data using a homomorphic algorithm such that data is manipulated while still being in cipher form.

Problem statement

Amid the awakening of covid-19, technology has enabled work progression has individuals were not required to work ideally in the workplace but remotely. In the local villages, specialized experts cannot be realized due to reduced medical practitioners employed by the government. Therefore telehealth has been learned can work for them.

Additionally, most patient records are kept as physical files, anticipating that they can be tampered with by the trusted insider hence denting the reputation of individuals and companies.

Problem Statement

In most hospital sets, it is evident that there are no adequate services offered simply because of reduced facilities and experienced professionals. Telehealth consultation can be done online simply because the model we are proposing is a model that is likely to enhance the security of our data. Additionally, the model will improve the safety of the data we transmit.

The Health centre’s lab test result takes a lot of time; hence they are not real-time. Therefore this takes time as either labs or lab technologists are limited in number, thus introducing a delay in the medical setup of a particular facility. It should be noted that telehealth has been integrated into many developing sectors to enhance their health in their various scopes and capacities.

Proposed solution

In articulating the problem addressed in 1.2, we have developed a fog computing-based telehealth system that uses secure multi-party computation and homomorphic encryption. The reduced number of medical practitioners will facilitate better service as some appointments can be booked with the professor, and procedures are performed on the specific data.

The issue of the reduced number of laboratories and lab technologists, telehealth can solve that as the specimen is sent via any means of transport for analysis. This makes sure that appropriate service has been acquired at all times.

Medical diagnosis is relayed through the made medical monitoring management system where individuals can access their data if need be. The data sent to fog servers are prominent encrypted and secure, and free from digital adversaries interested in intercepting our data. Therefore, the novel model will institute our data security as it is ever encrypted and our fog servers are ever untrusted [2]. The issues of delay in transmission of our results fog computing enhance decentralisation of data at a different geographical location. This ensures effective and efficient data retrieval at times and in the same implementations.

 Screenshot illustrating fog computing. 

Research aims and objectives

• To design a secure and free model from the digital assailant by applying a homomorphic encryption scheme based on data aggregation for medical monitoring systems.
• To engage fog-based computing to enhance our data transmission rate, which would otherwise bring delay in urgent medical data retrieved from the system.

Research question 

Research questions are kinds of queries that researchers need to articulate before they try to implement so that we can achieve the objectives set above in place[4]. The following are questions that include and are not limited to the following;

• What are network vulnerabilities associated with fog computing?.
• Apart from homomorphic encryption algorithm, do we have any other encryption algorithm which can work with fog servers with a similar or higher level of security?.
• Apart from telehealth, do we have any other sector where the remote procedure is offered?.

Research justification

Research work has been carried out to help support our research domain. Telehealth has been implemented using various algorithms. These algorithms, though working, exhibit some challenges which are addressed by this research. Let us focus, for instance, on a genetic algorithm that employs machine learning to predict certain kinds of diseases that individual’s exhibit.

The development of sensor-based systematic takes spaces from the body patient’s body. The efficiency of such a system has been questionable depending on the logic being used to make everything work. Therefore, security has become the main concern from the word go to the finality of all the algorithms being used; this research mitigates almost all the challenges exhibited by all other algorithms [8]. It should be noted that fog servers where our data is stored are in an untrusted environment.

Proposed Solution

The merit of using fog-computation, which is middleware in cloud computing services, is there is no downtime as fog servers as distributed in different geographical locations, and information is tapped from the nearest fog server; this means that there is no Jitter in packets sent across the network.

The security is also enhanced all along; even data manipulation involves manipulating data that is already encrypted. Therefore, there is no single data that data might be intercepted or impersonated by the digital assailant. This guarantees the security of the data.

Limitation of the research.

The research only uses only homomorphic encryption (HE)[5]. Even though it tries to eliminate some of the challenges exhibited by another algorithm, the fog server is located in an untrusted environment, which predisposes our data security.

Institutions that have implemented telehealth using another mechanism apart from homomorphic encryption algorithms and fog computing need to consider consent and ethical consideration.

The research covers only a tiny narrow domain, as another encryption algorithm will not be compared in this kind of research.

Summary

The key point of this part has been established by setting aims and objectives of the research. Whereby we want to design a secure free model from the digital assailant by applying a homomorphic encryption scheme based on data aggregation for medical monitoring systems. Moreover, having established the problems we need to canvas we need to find what other scholar have done, achieved and what we can do to make our model better. Therefore call for through literature review and theory which needs to be discussed in the next chapter of this paper.

Outcomes

In this section we intends to carry literature review so that we can be able to Marshall, the problems critical work of what others have done and addressed our problem as discussed in chapter one of this research.

Gap identification after literature review.

Summaries of the reviews.

Literature is nothing other than a critical appraisal of other people’s work. It is good to acknowledge them if you have used a particular work concerning them. In this research, the following analysis has been used to make this research a success. It is evident to evident that it plagiarized work if you don’t ensure that you reference whoever work has been used and acknowledge them well.

Scholars such as Bos et al.[25] carried out a very particular model using a lattice-based homomorphic cryptosystem to prediagnosis several medical conditions.Infact,with certainty and the goodness of the model is the ability to carry out all these tasks when it was required with a  high level of accuracy and precision. At this point, the data was being stored in the clouds as opposed to the way we’re keeping our data in fog servers. This model proved some effectiveness in handling some of the singular tasks, but it could not demonstrate the data entirely was immune from attacks [36][16]. The model suffered from a lack of confidentiality, especially from the trusted insider. In this model, it was a great effort of the aforementioned scholar to ensure that a lot of medical conditions have been diagnosed and the result conveyed to their corresponding destination without a lot of problems but the issue of delay and attack by an internal individual was not addressed at this point[19].

Research Aims and Objectives

Though scholars devoted user effort to e their attempt to ensure the privacy of the medical records, their server stored on the clouds is an external adversary as the internet is the home of the digital assailant at all times. This anticipates and means that for the upcoming researcher, they had to identify the gap and find what could be the best mechanism to improve the confidentiality of the medical records, which is required to be done in advance.

Moreover, when the scholars realized that the mode could achieve privacy of the medical records and the diagnosis of the disease, they now focused on improving the confidentiality of the data. In a bid to consolidate the secrecy of the model, Bos et al.[25] designed a classifier called job-privacy serving classify, which was anchored on three practical algorithms (decision tree algorithm, naïve Bayes, and hyperplane). The classifier acted as a medic and was secure and free from digital adversaries n a stable cryptosystem called a paillier cryptosystem.

Other opportunistic scholars who seized the opportunity in the prediscussed model could improve the small functionalities of the models to make it better and better. For instance, Wahu et al. [7] designed a privacy-preserving model which enhanced and improved the security of the prediagnosed records which were missing the model. It is worthy to note that many concerns have been enhancing the security and confidentiality of medical records; it should be noted that real-time retrieval of documents is of significant importance.

It should be noted that the two medical models have not focused on the availability of the data and retrieval rate. This is a gap and needs to be articulated. Fog computing provides this platform as an architect that offers real-time delivery of medical records when required.

Medical records are very scant as they contain private and personal data requiring privacy and confidentiality [29]. Availability of the same has become very advent and forthcoming. Therefore all models developed should exhibit some mechanism of real-timeliness and some component of urgency. Some of the renowned scholars applied opportunistic computing in their research careers, which has  proven effective in enhancing the security of the data. It is worthwhile noting that security of data once enhanced all data will work in the model so that accurate results can be enhanced at all times.

Lu et al. [13] proposed an opportunistic framework primarily meant to assist people in the village by setting up emergency healthcare. The work of this model was only to achieve privacy and the reliability of the medical records. This kind of model is referred to as an opportunistic computing framework.

However, though the model enhances confidentiality, privacy, and availability, of data availability framework suffers from the ability to determine how long a specific computing device will take time to process the availability of a particular service. It is worth noting that real-timeliness is a constant factor in every model. The discussed model also suffers from a lack of this in their model. Therefore the urgency of the data determines which algorithm and architecture to be used in the data processing approach.

Research Questions

Another key concern that should be critically addressed to finality to enhance the security of the data is through the anonymization process. This process is involved concealing a certain amount of data at all times. This, in turn, will hide the identity of our customers at all times. Some of the standard particulars which are hidden from the public include and are not limited to the following; usernames and their corresponding postal addresses. But in fact, this is not a big deal for a digital assailant to retrieve this information. This was a cheap mechanism devised and proposed by. Machanavajjhala et al.[22].The digital assailant can perform backtracking, identify the most appropriate tool, get the appropriate data, and exploit the information they need to acquire. Data is now being stored in the database; digital assailants use denoising techniques to identify the information which is being blurred during the process as data is transmitted from one point to another. 

Ideally, what ought to be done is to ensure are synchronicities of data being transferred. Notionally, due to frequent attacks on this method, it is not advisable to use this traditional method of data concealing, t; this simply because the data will involve user data being circumnavigated at all times to enhance the security of the data at all times. This mechanism is susceptible to attack while sending our data packets from one destination point to another. Therefore, it can be identified that anonymization cannot be, will not be, and should no  be a suitable mechanism for protecting patient records when telehealth is used to offer services in a hospital setup.

Privacy-preserving models have been established in many sectors such that the security of data can be established in a very coherent manner. The various framework has been designed to enhance frameworks have been designed to enhance confidentiality and security data at all times. For instance, a framework to improve the security of the smart city. A smart city is an implementation of the internet of things. The security of data in this kind of model or setup requires one to implement a framework that shields data sent from one point to another. Smart cities are easily susceptible to attack due to the increased surface of attacks. Ideally, if not well protected and immune from attack data which are transmitted across various institution up a smart city, where various number integrated system are involved there will be vast and severe loss of data. Attackers take this as an advantage to ensure work has been established fully at all times.

Lett et al.[17] devised a model framework that could be used to enhance the security of data as it is being transmitted from one point to another. The scholar used very intuitively, knowledgeable improving data. The model used in this context is divided into three privacy levels. That is from level 1to 3. Classifying data in level ensured that discount data from their security could be enhanced and deliberately performed. Audio, video and plain document security were enhanced depending on the secrecy the file exhibited for the transmission of that particular.

Research Justification

Therefore different levels of secrecy depend on the level of confidentiality of data to be transmitted. This should improve the security of encrypted data being sent across the network. It should be noted that the model had a mechanism of maintaining privacy. What about the readiness of the retrieval of information. Therefore this model suffered from the ability to issue data when required. Therefore the concept of arises in this framework as well.

The fog latency has become another aspect of concern in fog computing; therefore, a team of experts has devised a mechanism to shield all elements of network latency to expedite the required speed. Gaus et al. [19] developed a model that used fog-based data aggregation architecture. The model aimed to denoise the communication channel and achieve a deterministic noise level with zero decibels, reducing the latency in packet transmission. This mechanism improved the response time of packets sent across the network from one point to another.

The sensor-based system enhances also requires a high level of security at all, all de; all which are used in the network, which is digital assailant prone, require a high level of security in enhancing their day-to-day operations. We compare the work of Ara et al.[24] it implemented a data aggregation approach that guaranteed the protection of the user data.

Research scholars such as Zhou et al.[35] devised a mechanism of such as secure multiparty computation. The security mechanism enhanced the paradigms of packet transmission. This security-security enhanced the city by effectively ensuring maximum mobility, which necessitated data movement from one point to another. These paradigms also involved another component of security-enhanced, which include and is not limited to homomorphic encryption standards and, finally, trap door function, which enhanced the security of the data on the already designed security paradigms at all times.

Research Gap

Research gap refers to that part of many research that has not been yet implemented and yet needs to be implement to ensure a certain model works optimumly.This gap appears to be the weakness of those research or system implemented. In our research it was evident that one of the recurring theme, is the aspect of security on all research work. For the analysis which has been made to many review journals theory has demonstrated many research have articulated effectively security and availability of the data will in transit. However though security have been achieved, it as been identified it is not absolute. A times digital assailant could launch an attack. Additionally other framework focused much on security without considering, rate of the execution of the data they are enhancing security. What the delay time for the data being retrieved from the cloud server or database. All these are niche which our model we implement by applying homomrorphic encryption and fog server. To mitigate the issue of delay we shall engage fog server which operate in real-time in data transmission and retrieval and homomorphic encryption which ensure that security is enhanced as data is manipulated will being encrypted.

Limitation of the Research

Additionally, secure multiparty computation will enhance the effect of security as individual will have private key as they as compute privately before exchange key. This simulate the concept of key cryptography.

List of work covered by different scholars. 

Author	Issued achieved	Approach
Bos et al.	Security and availability	Lattice-based homomorphic Cryptosystem
Bos et al	Security and integrity	Hyperplane, Naïve Bayes and decision tree
Zhou et al.	security	Security multi-party computation
Gaus	security	Secure multiparty computation and homomorphic encryption
Lett et al	Security and integrity	Secure computation computation by dividing data into 3 security levels
Machanavajjhala	security	Public key cryptography(AES)

Summary

This is most important of part of the research as it sheds light what has been established or not. Therefore having established the gap and identified areas and environment we want to improve, we have identified medical field and the cloud environment using fog-based computing middleware. We intend to fill the gap of delay and security through the application of fog-based computing and homomorphic encryption. To establish this and bring this to reality we design our models as defined in chapter 3 below research methodologies and specifications.

Research methods are part of research that enhances the realization of what has been discussed in (i) and (ii). Critical statistical methodologies are applied to ensure appropriate data has been achieved. This will lead to reporting achieving some of the goals set and the problem identified to be solved by the research.

Research specifications.

Research specifications are important edicts in research as they demonstrate what the research is trying to achieve and fine-tune by establishing justification of the same. A Times they are referred to as deliverables simply because they are implementable in the project. If specifications are not addressed to finality, it simply means that research has not attained its required objectives. Our telehealth model requires us to address the challenges proposed in the problem domain. They include and are not limited to the following.

(i)The proposed model should ensure that the data transmitted across the network is secure .The fog server should immune from interception by the digital assailant [14]. The digital assailant is to intercept data involved in the aspect data so that they can claim the originality of the data. The proposed model mitigates these issues by ensuring appropriate homomorphic encryption before storing data on the cloud server.

(ii)The proposed medical scheme should enhance accurate and faster data retrieval. This reduces unnecessary transport costs as all results can be transported from one point to another. This solves the problem addressed earlier on where we have reduced the number of medical practitioners in remote areas in societies where there is less development.

(iii)The model proposed should exhibit a high level of security to our medical system where individual data at no time should be viewed by individuals who are not intended to achieve that[16][11]. For instance, it is worthwhile to note other encryption standard algorithm is subject to attack as the older encryption algorithm could be attacked or infringed by a digital assailant with the help of a trusted insider. Moreover, our model makes use of homomorphic encryption, which prevent event trusted insider from leaking the data at their disposal. Every actor cannot be able to manipulate the data as the data from the fog server is always encrypted.

(iv)The proposed model should deal with finality on the issues of delay, which has not been addressed by several models discussed in the literature review [15]. What should be understood is that everyone who worked researched and verified it lacks the aspiring delivery of data. The response time was not synchronized in the algorithms and schemes used as they focused on the security and confidentiality of the data.

Summary

Proposed methodology 

Methodology is the procedure which needs to be articulated in order to address challenges that has been identified in the research gap. The methodology we intend to adopt in this research is agile methodology in implementing this work. Appropriate mechanism is implemented to enhance the security of records transmitted between the patient and the doctors. The agile methodology makes use of homomorphic encryption and secure multiparty computation to address the challenges of real-time and security of our data in transit.

Each and every part of implementation will be carried out sequential in modular form so that we can enhance appropriate security of our packets through testing and addressing the appropriate challenges for our data packets. In our implementation the issues of concerns and attributes of fog-based computing make it salient for our implementation. Additionally, the security of patient data is enhanced by using secure multi-party computation as per our model’s standards. Traditional methods of data enhancement had a lot of challenges that the proposed model is addressing. For instance, when you check models and schemes such as anonymity, encryption standards could be intercepted easily by the digital assailant with less effort.

 

Self-serviced medical diagnosis for our model. 

Secure multiparty computation

This mechanism will enable the secure transfer of patient data from one point source to another. This is the mechanism in which our data, the data we are transferring packets from one destination[19][21]. Data will be required and enhanced through the safest tool. Ideally, when patient data has been computed using homomorphic encryption, secure multiparty computation is done, and improving the same is done through the most appropriate mechanism for enhancing the security of the data we are trying to establish.

In this context, computation will be done by different separate parties to generate secure data which cannot be intercepted by the cheater or an attacker who may be interested in eavesdropping on the information we have at our disposal for transmission of our data. This mechanism will enhance our data privacy as we understand our environment for our fog server is not secure at all.

Implementation procedure in our research to enhance security.

The actor’s research will be doctors and patients whose patients interact; therefore, they request the service while the other one provides the services required during the implementation process at all times. It is worth noting that secure multiparty is one of the safest methods as data for manipulation is private and c, computation is keyless. Therefore, using two, participants can be referred to us as secure two-party computation.

Argumentatively, to identify the computation result where the doctor and their patient interact privately to avoid digital assailants, the following will be the calculation results between the two involved parties[12][16]. When the two parties work together, assuming they have keys m₁ and m_2.We obtain the development of the computation by finding a function that the two parties establish. This will look like Y=f(m₁₊m₂).

This method of adoption has been used even there before and has demonstrated it is very eminent in ensuring the security of very critical data. When this computation has been done, objectives point number one will have been achieved in totally, has the safety of the data will have been performed at all times in enhancing the protection of the data we have at hand.

Literature Review

Homomorphic encryption

This is the encryption algorithm that we shall use to shield our data so that digital assailants will not have any opportunity to intercept our data correctly and appropriately.

First, this mechanism requires one to identify which kind of data is being encrypted so that it can remain anonymous online and prevent digital assailants from taking place. Ideally, it is by default patient records, and individual hospital records are the word to be encrypted in advance before being stored in our fog-based servers[19]. The following approach will be undertaken to encrypt a batch of data from one point to the required destinations.

Steps one:

Choose appropriate tactic, techniques procedure security entity such as generate a specific key

Choose an appropriate cryptographic algorithm that will be used to hash your data packets as they are required to be transferred from one point destination within the confines of the perimeters of the network. This hashing function will be the one that will be involved in computing the hash of the plain text we intend to send across the network. The plain text will probably be the medical report and data from the hospital setup.

Step 2.

All the data which have been generated is just shared with all involved parties in the network system environment. Whatever we are sharing is nothing other than the parameters which were chosen in step one mentioned above.

Step 3.

The fog server needs to establish communication by generating a private key p and a public key computed by the functions Y=K_P.

Step 4.

Now we  all the required keys to be authenticated in our available network design. Therefore tactic, techniques procedure assigns the private keys to all devices present in the network. Assuming the private key in this context will include the following parameters p={ s_i,j, v_i,j, u_i,j, d_i,j }.Simply because this kind of network is exceptional, the user is identified with a unique identifier that identifies himself in the network and the fog server(FS)[26][31]. This is done uniquely and the same so that we can have synchronized participant’s data in the network and, in this context, the doctor and the patient.

Step 5

The availability network, intelligent enough for its surroundings, chooses private and public keys to be used in the computation.

Step 6

Deploy some present threshold to already encrypted data (Medical report data), which is ready to be sent across the network by the tactic, techniques procedure security. This is the data sent to the fog server for storage.

Design architecture diagrams

Flowchart of fog server

Usecase diagrams

Usecase actually represents the interaction between several actors and the system in question. In our use-case demonstration, it is very effective and easy to understand what are some of the actions that an actor in the system can perform. In this, we are the ones we are intending enhance security through our homomorphic encryption and secure multiparty system.

Usecase No: Use_ 001. 

The usecase given above has got 3 actors on the system who can actually demonstrate the interaction between the user and the system.

Description of the usecase.

Name	Usecase No:user_001.
Actors	Doctor, patient, System Administrator.
Main flow	First we must access the index page using the provided url. ü Enter the required password and email. ü Check the doctor available, time and their specialization ü Book appointment with specialist and wait for the reply from the portal. ü Check out reply from the specialist after sometimes. ü Log out
Alternate flow	ü Log in into the system ü Raise exceptions once incorrect credentials have been issued.
Pre-conditions	ü Approach the dashboard. ü Enter Log in credentials ü Incorrect credentials show a message dialog to ask the user to enter valid credentials. ü Correct credentials authenticate to access various features of the application.
Post conditions	ü Book appointment. ü Initiate video call, for cross examination by the doctor. ü Check result from the system ü Log out
Non-functional requirements	These are conditions that cannot be implemented but the making service effective at all times. Some of these non-functional trait include and not limited to the following  security, availability, reliability, and the ability to mask failure through catching exceptions.

Sequence diagram

 

Sequence diagram for retrieval of information on the fog-server.

Project milestone

The milestone of the project matters a lot it is good to identify the time spent in the entire process, in handling various parts and aspects of the project. In this project, we will apply the use of a Gantt chart in order to display various milestones of the work we have already carried out.

Conclusion

This is one of the important parts of research as it provides a full depiction of what has been enhanced throughout the research. Great projections are made from the onset by instating various domains undertaken in the research. It is worthwhile to note that data appropriation is made by enhancing the security of the data by the available description which already has been set.

Having acquired all the required prerequisites to implement all that is required in the business, it is worth noting that the medical report will be secure from the hands of a digital assailant. The total effectiveness of the algorithm will also encompass individuals to take the critical approaches in ensuring the security of the data. Either in transit(during fog computing) or the fog servers. In totality, we need to ensure that appropriate security hence been enhanced for our security data. Medical records are very sensitive; hence critical security mechanisms should set the stage for our data.

 

The screenshot above shows how the patient record is processed before being stored in the medical database. In our case, we intend to process to store this data in our fog service. This is simply because it will enhance the utility of this data at all times.

Homomorphic encryption has been used to enhance the security of the available data hence promoting telehealth technology at all levels of hospitals and different nations at all

From the research, it can be ascertained that data secure encrypt certain date joint computation to a keyless user in the system. In fact, what should this mechanism have also helped enhance the security of our data?. Fog server should be implemented at any cost as our self-preserving privacy will be highly enhanced by our system[30]. This mechanism ensures faster retrieval of information by application of fog server. This method should be enhanced to all institution that needs to develop this kind of information at all times. We ought to understand that through the fog server is in untrusted environments, data stored are very secure at all times.

References 

[1] Georgios, L., Kerstin, S., & Theofylaktos, A. (2019). Internet of things in the context of industry 4.0: An overview. 

[2] Peng, Z. (2019).The danger of using fully homomorphic encryption: A look at Microsoft SEAL. arXiv preprint arXiv:1906.07127. 

[3] Al Badawi, A., Veeravalli, B., Lin, J., Xiao, N., Kazuaki, M., & Mi, A. K. M. (2020). Multi-GPU design and performance evaluation of homomorphic encryption on GPU clusters. IEEE Transactions on Parallel and Distributed Systems, 32(2), 379-391. 

[4] Roy, S. S., Turan, F., Jarvinen, K., Vercauteren, F., & Verbauwhede, I. (2019, February). Abased-bbasedbasederformanhbasedrformance-base based for morphic computing on encrypted data. In 2019 IEEE International symposium on high-performance computer architecture (HPCA) (pp. 387-398). IEEE. 

[5] Lu, Y., Zhou, T., Tian, Y., Zhu, S., & Li, J. (2020). Web-Based privacy-preserving multicenter medical data analysis tools via threshold homomorphic encryption: design and development study—Journal of medical Internet research, 22(12), e22555. 

[6] Chialva, D., & Dooms, A. (2018). Conditionals in homomorphic encryption and machine learning applications. arXiv preprint arXiv:1810.12380. 

[7] Wahu H., Muyeen, S. M., Mahmoudi, A., Benbouzid, M., Gharehpetian, G. B., & Kabalci, E. An Agile Black-Out Detection and Response Paradigm in Smart Grids Incorporating IoT-Oriented Initiatives and Fog-Computing Platform. 

[8] Anawar, M. R., Wang, S., Azam Zia, M., Jadoon, A. K., Akram, U., & Raza, S. (2018). Fog computing: An overview of big IoT data analytics. Wireless Communications and Mobile Computing, 2018. 

[9] Naha, R. K., & Garg, S. (2021). Multi-criteria–based Dynamic User Behaviour–aware Resource Allocation in Fog Computing. ACM Transactions on Internet of Things, 2(1), 1-31. 

[10] Martinez, I., Hafid, A. S., & Jarray, A. (2020). Design, resource management, and evaluation of fog computing systems: a survey. IEEE Internet of Things Journal, 8(4), 2494-2516.

[11] Kim, M., Song, Y., Wang, S., Xia, Y., & Jiang, X. (2018). Secure logistic regression based on homomorphic encryption: Design and evaluation. JMIR medical informatics, 6(2), e8805.

[12] Al Badawi, A., Polyakov, Y., Aung, K. M. M., Veeravalli, B., & Rohloff, K. (2019). Implementation and performance evaluation of RNS variants of the BFV homomorphic encryption scheme. IEEE Transactions on Emerging Topics in Computing, 9(2), 941-956.

[13] Lu, J. H., Han, K., Kim, A., Kim, M., & Song, Y. (2018, April). Bootstrapping for approximate homomorphic encryption. In Annual International Conference on the Theory and Applications of Cryptographic Techniques (pp. 360-384). Springer, Cham.

[14] Viand, A., Jattke, P., & Hithnawi, A. (2021, May). SoK: Fully homomorphic encryption compilers. In 2021 IEEE Symposium on Security and Privacy (SP) (pp. 1092-1108). IEEE.

[15] Al Badawi, A., Veeravalli, B., Lin, J., Xiao, N., Kazuaki, M., & Mi, A. K. M. (2020). Multi-GPU design and performance evaluation of homomorphic encryption on GPU clusters. IEEE Transactions on Parallel and Distributed Systems, 32(2), 379-391.

[16] Aloufi, A., Hu, P., Wong, H. W., & Chow, S. S. (2019). Blindfolded evaluation of random forests with multi-key homomorphic encryption. IEEE Transactions on Dependable and Secure Computing, 18(4), 1821-1835.

[17] Lett, M., Alani, M. M., Al-Jumeily, D., Baker, T., Mustafina, J., Hussain, A., & Aljaaf, A. J. (2019). A systematic review on the status and progress of homomorphic encryption technologies. Journal of Information Security and Applications, 48, 102362.

[18] Genise, N., Gentry, C., Halevi, S., Li, B., & Micciancio, D. (2019, December). Homomorphic encryption for finite automata. In International Conference on the Theory and Application of Cryptology and Information Security (pp. 473-502). Springer, Cham.

[19] Gaus, C., Tao, M., Zhang, J., Hong, X., & Yuan, R. (2018). Blockchain based credibility verification method for IoT entities. Security and Communication Networks, 2018.

[20] Clark, C., Yatskar, M., & Zettlemoyer, L. (2019). Don’t take the easy way out: ensemble based methods for avoiding known dataset biases. arXiv preprint arXiv:1909.03683.

[21] Alli, A. A., & Alam, M. M. (2019). SecOFF-FCIoT: Machine learning based secure offloading in Fog-Cloud of things for smart city applications. Internet of Things, 7, 100070.

[22] Machanavajjhala, Z., Xu, L., Chen, L., Zhao, X., Lu, Y., & Shi, W. (2018). CoC: A unified distributed ledger based supply chain management system. Journal of Computer Science and Technology, 33(2), 237-248. 

[23] Melnikovas, A. (2018). Towards an explicit research methodology: Adapting research onion model for futures studies. Journal of Futures Studies, 23(2), 29-44.

[24] Ara, D., Liao, X., Xiang, T., Wu, J., & Le, J. (2020). Privacy-preserving self-serviced medical diagnosis scheme based on secure multi-party computation. Computers & Security, 90, 101701.

[25] Bos, C., Zhu, L., Xu, C., & Lu, R. (2018). PPDP: An efficient and privacy-preserving disease prediction scheme in cloud-based e-Healthcare system. Future Generation Computer Systems, 79, 16-25.

[26] Guo, W., Shao, J., Lu, R., Liu, Y., & Ghorbani, A. A. (2018). A privacy-preserving online medical prediagnosis scheme for cloud environment. IEEE Access, 6, 48946-48957.

[27] Meshram, C., Ibrahim, R. W., Obaidat, M. S., Sadoun, B., Meshram, S. G., & Tembhurne, J. V. (2021). An effective mobile-healthcare emerging emergency medical system using conformable chaotic maps. Soft Computing, 25(14), 8905-8920.

[28] Kuru, K. (2021). Planning the future of smart cities with swarms of fully autonomous unmanned aerial vehicles using a novel framework. IEEE Access, 9, 6571-6595.

[29] Dewangan, K., & Mishra, M. (2018). Internet of things for healthcare: A review. International Journal of Advanced in Management, Technology and Engineering Sciences, 8(3), 526-534.

[30] Chen, W., Chen, Z., & Cui, F. (2019). Adaptive routing optimization algorithm in community-oriented opportunistic networks for mobile health. Sensors, 19(8), 1876.

[31] Knott, B., Venkataraman, S., Hannun, A., Sengupta, S., Ibrahim, M., & van der Maaten, L. (2021). Crypten: Secure multi-party computation meets machine learning. Advances in Neural Information Processing Systems, 34.

[32] Choi, J. I., & Butler, K. R. (2019). Secure multiparty computation and trusted hardware: Examining adoption challenges and opportunities. Security and Communication Networks, 2019.

[33] Ahmad, A., Webb, J., Desouza, K. C., & Boorman, J. (2019). Strategically-motivated advanced persistent threat: Definition, process, tactics and a disinformation model of counterattack. Computers & Security, 86, 402-418.

[34] Liu, J. N., Weng, J., Yang, A., Chen, Y., & Lin, X. (2019). Enabling efficient and privacy-preserving aggregation communication and function query for fog computing-based smart grid. IEEE Transactions on Smart Grid, 11(1), 247-257.

[35] Zhou, J., Wu, L., Wang, H., Choo, K. K. R., & He, D. (2020). An efficient and privacy-preserving outsourced support vector machine training for internet of medical things. IEEE Internet of Things Journal, 8(1), 458-473.

[36] Shan, J., Lin, Y., & Zhu, X. (2020, March). A new range noise perturbation method based on privacy preserving data mining. In 2020 IEEE International Conference on Artificial Intelligence and Information Systems (ICAIIS) (pp. 131-136). IEEE.