KryptosChain—A Blockchain-Inspired, AI-Combined, DNA-Encrypted Secure Information Exchange Scheme

Round 1

Reviewer 1 Report

This paper uses Huffman Coding fundamentals due its feature of as signing variable codes and smaller codes for more frequently occurring symbol. This paper is limited contributions, but extensive results and its analysis getting attention on it. However, this topic is old and overused for this problem in the literature. But, the analysis results in matter, this paper suggesting following comments for revising to improve the paper quality.

1. This paper can be improved in Theoretical aspect. This paper is poor in theoretical analysis.

2. The reasons for achieving the superior performance over the previous or existing work must be provided in the paper.

3. For which applications this work is more suitable? Computational complexity also considered to choose the best applications for this work.

4. What are the implementation challenges impose when applied this model to real time applications?

5. Provide the units for Y-axes, for all the plots in the paper.

6. There are several statistical metrics in the literature, why was the authors considered only four metrics such as accuracy, precision, F-score, and recall. The authors must evaluate the multiple correlation metrics available in the statistical analysis models.

7. Provide the citations for the datasets used in the paper. In addition, the source code of this paper can be provided through GITHUB link, to check the reliability of implementation.

8. Provide the clear description for Figure 10.

9. Provide the pseudo codes for the algorithms mentioned or proposed in the paper.

10. Elaborate the prime highlighted mentioned in the introduction of the paper.

11. Provide the motivation of the paper through an illustrative example.

12. Half of the abstract contain the general information. It is avoided. Provide the proposed work and its contributions, pitfalls, and summary on results in 125 to 250 words. Avoid the general discussion which is available in the introduction. Follow the same for Conclusion as well. The discussion can be moved to previous sections from the conclusion.

Author Response

Comment 1: This paper can be improved in Theoretical aspect. This paper is poor in theoretical analysis.

Response: Thank you for your suggestion. The overall theoretical analysis has been added in Introduction section which are highlighted in yellow and marked in track changes scattered in Pages 1 - 3. Section 5 from Page 23 also provides the detailed analysis of the proposed model.

Comment 2: The reasons for achieving the superior performance over the previous or existing work must be provided in the paper.

Response: Thank you for your suggestion. Sections 5.1.5, 5.2.3, 5.3 provides the tables and paragraphs to put forward the necessary comparison.

Comment 3. For which applications this work is more suitable? Computational complexity also considered to choose the best applications for this work.

Response: The proposed work is a secure information exchange scheme between 2 parties. 

Comment 4. What are the implementation challenges imposed when applied this model to real time applications?

Response: The time to generate the DNA based Huffman Code depends on the number of n distinct symbols with a complexity of O(n log n). Currently this paper considers only text messages and other formats of information will be tried upon in our future endeavors.  The Blockchain inspired exchange mechanism is time taking because of the RSA algorithm used to generate the public and private keys. The actual implementation of SHA-256 is complex because of the steps involved such as Padding, Append Length, Division of data into blocks and initialization of chaining variables. Also the 6 phases make the block generation and exchange tedious. The challenges faced during the AI based intrusion detection is false alarm in the DoS attack and slow response time for the R2L attack.

Comment 5. Provide the units for Y-axes, for all the plots in the paper.

Response: The figures have been updated.

Comment 6. There are several statistical metrics in the literature, why was the authors considered only four metrics such as accuracy, precision, F-score, and recall. The authors must evaluate the multiple correlation metrics available in the statistical analysis models.

Response: Two more parameters - Time to train and Sensitivity has also been added in Table 16. Other parameters will be studied and analyzed in the future scope of our work.

Comment 7. Provide the citations for the datasets used in the paper. In addition, the source code of this paper can be provided through GITHUB link, to check the reliability of implementation.

Response: The datasets citations are provided in the Reference list number 52 - 53. As of now only the detailed flowcharts related to the proposed scheme are provided in the paper. The source codes are not shared in public platform as the work is unpublished and under review.

Comment 8. Provide the clear description for Figure 10.

Response: The description has been added in Page 14 towards the bottom of the page.

Comment 9: Provide the pseudo codes for the algorithms mentioned or proposed in the paper.

Response: Detailed flowcharts has been provided for all the proposals. Wherever possible the pseudo code has also been added.

Comment 10: Elaborate the prime highlighted mentioned in the introduction of the paper.

Response: The elaborations have been added it page 2 - 3.

Comment 11: Provide the motivation of the paper through an illustrative example.

Response: An example scenario has been added in the introduction section in Page 1.

Comment 12. Half of the abstract contain the general information. It is avoided. Provide the proposed work and its contributions, pitfalls, and summary on results in 125 to 250 words. Avoid the general discussion which is available in the introduction. Follow the same for Conclusion as well. The discussion can be moved to previous sections from the conclusion.

Response: The abstract and conclusion have been updated as per your valuable suggestions

Reviewer 2 Report

In this paper, the authors proposed a blockchain and AI-based combinatorial framework along with DNA encryption to secure the data exchange between sender and receiver against various cryptographic attacks. However, the paper has the following major concerns that can affect the novelty of the paper.

1) The Introduction section does not represent the main motivation of the proposed framework. As they have not explained the research gaps of the conventional schemes that failed to provide security in the system. And how the integration of blockchain, AI, and DNA encryption is going to strengthen the data security of the system.

2) Subsection 2.6 is not completed which impacts the understanding and quality of the manuscript. There is no comparative analysis performed for the analysis of existing AI-based cryptographic schemes, instead they have directly started the proposed methodology making the paper inappropriate.

3) The authors have represented the actual information exchange process in Phase 6 of the proposed cipher information transmission scheme with the help of equations 1 and 2. But,  the equations need to be clarified as they have not defined variables. Due to which, it is  difficult to understand the relation between equations. Moreover, Figure 7 needs to be explained elaborately as it depicts the actual information exchange process using blockchain.

4) Figure 7 depicts the usage of blockchain to secure the information exchange process. So, there must be involvement of a consensus mechanism for securing the information. But, authors have not mentioned any consensus mechanism which can affect the data security of the system making it vulnerable to various malicious attacks.

5) In subsection 3.3, the authors have proposed an intrusion detection scheme to secure  the communication between the intended sender and receiver. But, the main focus of the paper, i.e., security, cannot be fulfilled by implementing the AI-based model for intrusion detection. As malicious attackers can easily target the data stored via an AI-based model  exposing it to various security threats.

6) In subsection 5.2, the authors have highlighted the achievement of access control while the attainment of principles of security is achieved. But, it is not sufficient to prove the mechanism of access control in the proposed scheme. As they did not consider any consensus mechanism due to which adversaries can easily access the data to further manipulate it disrupting the security of the network.

7) In this paper, the authors mainly focus on providing security in the proposed scheme. But they have not performed any security analysis on the proposed scheme further affecting the novelty of the paper. Also, in conclusion, they discuss the reduced computational complexity of the proposed scheme. But how they are going to reduce it? As, usage of AI-based model require relatively high computation time to process huge amount  of data and also incur huge data storage cost further making it difficult to implement it in real-time scenario.

Author Response

Comment 1: The Introduction section does not represent the main motivation of the proposed framework. As they have not explained the research gaps of the conventional schemes that failed to provide security in the system. And how the integration of blockchain, AI, and DNA encryption is going to strengthen the data security of the system.

Response: The introduction has been updated taking into account the valuable comments of the reviewers.

Comment 2: Subsection 2.6 is not completed which impacts the understanding and quality of the manuscript. There is no comparative analysis performed for the analysis of existing AI-based cryptographic schemes, instead they have directly started the proposed methodology making the paper inappropriate.

Response: The subsection has been deleted. The contents has been discussed in the existing work itself.

Comment 3: The authors have represented the actual information exchange process in Phase 6 of the proposed cipher information transmission scheme with the help of equations 1 and 2. But,  the equations need to be clarified as they have not defined variables. Due to which, it is  difficult to understand the relation between equations. Moreover, Figure 7 needs to be explained elaborately as it depicts the actual information exchange process using blockchain.

Response: All the variables have been defined before the beginning of section 3.2.1. Figure 7 explanation is also added in Pages 11-12.

Comment 4: Figure 7 depicts the usage of blockchain to secure the information exchange process. So, there must be involvement of a consensus mechanism for securing the information. But, authors have not mentioned any consensus mechanism which can affect the data security of the system making it vulnerable to various malicious attacks.

Response: The proposed model tries to put forward an information exchange scheme which is only blockchain inspired and a refinement on the Diffie - Hellman Exchange scheme by inculcating the fundamentals of blocking that exploits the concept of hashing. Real Blockchains mandate possession of cryptocurrency, deploying smart contracts and defining suitable consensus mechanism which is expensive, time taking and difficult to practically implement.

Comment 5: In subsection 3.3, the authors have proposed an intrusion detection scheme to secure  the communication between the intended sender and receiver. But, the main focus of the paper, i.e., security, cannot be fulfilled by implementing the AI-based model for intrusion detection. As malicious attackers can easily target the data stored via an AI-based model  exposing it to various security threats.

Response: The security of the communication is being achieve by a two step process. First by the DNA based Huffman coding assigns variable length DNA codes to the distinct symbols of the original message in-contrast to the traditional binary Huffman coding. Secondly the obtained ciphertext is transmitted by a blockchain inspired method which stores the hash of the previous block into the successive block. Any modification in the content with update the hash leading to a hash mismatch in the blocks ahead. Thus, tampering will be easily noticed. Also each user has to undergo the authentication process as imposed by the proposed scheme. Along with this as an additional feature, the paper proposes a continuous AI based IDS which will keep a tap on any intrusions and raise alarms at an early stage.

Comment 6: In subsection 5.2, the authors have highlighted the achievement of access control while the attainment of principles of security is achieved. But, it is not sufficient to prove the mechanism of access control in the proposed scheme. As they did not consider any consensus mechanism due to which adversaries can easily access the data to further manipulate it disrupting the security of the network.

Response : The paper showcases that there is trusted third party Kyrios involved in the initial phases of the KryptosChain who validates and authenticates the entities trying to gain access into it. Any unwanted entity can easily be debarred by Kyrios from accessing the system. Due to these features, the paper claims that access control is achieved.

Comment 7: In this paper, the authors mainly focus on providing security in the proposed scheme. But they have not performed any security analysis on the proposed scheme further affecting the novelty of the paper. Also, in conclusion, they discuss the reduced computational complexity of the proposed scheme. But how they are going to reduce it? As, usage of AI-based model require relatively high computation time to process huge amount  of data and also incur huge data storage cost further making it difficult to implement it in real-time scenario.

Response: The implementation constraints are definitely a hindrance but the prime emphasize of the paper is to ensure security and successfully transmit a message from the source to the intended receiver.

Round 2

Reviewer 1 Report

Thank you addressing all the suggested comments. This paper can be considered for publication in this journal.

Reviewer 2 Report

All comments are incorporated. No further comments.