An Effective Sharding Consensus Algorithm for Blockchain Systems
Round 1
Reviewer 1 Report
This paper proposes the Practical Byzantine Fault Tolerance (PBFT) as the intra-shard consensus and propose a new sharding consensus mechanism, which to minimize the mapping of the node assignment by combine jump consistent hash algorithm with signature Anchorhash. Furthermore, this paper solves the process of the cross-shard transaction and use the activity of nodes participating in intra-shard transactions as the criterion of shard reconfiguration and ensures the security of the blockchain system. This paper is structured correctly and useful contribution to solves the problem of uneven shard arithmetic power and insecure cross-shard transaction processing. These contributions were well formulated and the answers to them were correct and clear. I think that this paper is appropriate for publishing.
Author Response
Dear Reviewer:
Thank you for your letter and your suggestions on our manuscript. These comments are invaluable and help better notice the details in our articles and provide important guidance for subsequent research.
After we checked the entire article again, we found that there were many spelling problems in the article and many experimental descriptions and data require further elaboration. We have further optimized the content of the article according to the problems in the article to meet your standards
We would love to thank you for allowing us to resubmit a revised copy of the manuscript and we highly appreciate your time and consideration.
Best regards
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The paper presents a consensus algorithm for blockchain sharding. The proposed algorithm divides nodes into shards through an Achorhash hash function.
However, the authors do not present an attacker model and do not discuss the security of the protocol. In particular, the authors should discuss the consensus regarding the validity, agreement, and termination properties. Furthermore, authors must prove the liveness and safety properties of the proposed consensus.
The proof that the proposed mechanism meets Lamport's liveness and safety properties is essential for validating the proposal.
What is the attacker model that the proposal aims to resist? Does the proposed consensus resist Byzantine behavior?
It is not clear the need for an incentive (motivation) mechanism for blockchain based on Byzantine consensus.
The experiments should have focused on the probability of success of attacks that the consensus algorithm was supposed to resist. In this way, it is crucial to stand the attacker model previously and derive an analytical model for the attack success probability. Moreover, experiments should demonstrate the message and the time overhead for reaching the consensus. Finally, the authors should compare the proposal with previous work, such as ELASTICO.
There are minor typos. The authors should proofread the manuscript.
Author Response
Dear Reviewer:
Thank you for your letter and your suggestions on our manuscript. These comments are invaluable and help we recognize the shortcomings in the article and point us in the direction of improvement.
As you said, we did not present an attacker model and do not discuss the security of the protocol
Here are our responses and did not discuss the consensus regarding the validity, agreement, and termination properties very clearly. Meanwhile, We only explained Lamport's liveness and safety properties in very little space. We only do some analysis on the common problems of the sharding consensus algorithm and ignore the validity proof of our own algorithm. These are all our shortcomings.
Here are five points you listed:
Point 1: However, the authors do not present an attacker model and do not discuss the security of the protocol. In particular, the authors should discuss the consensus regarding the validity, agreement, and termination properties. Furthermore, authors must prove the liveness and safety properties of the proposed consensus.
Point 2: The proof that the proposed mechanism meets Lamport's liveness and safety properties is essential for validating the proposal..
Point 3: What is the attacker model that the proposal aims to resist? Does the proposed consensus resist Byzantine behavior?
Point 4: It is not clear the need for an incentive (motivation) mechanism for blockchain based on Byzantine consensus.
Point 5: The experiments should have focused on the probability of success of attacks that the consensus algorithm was supposed to resist. In this way, it is crucial to stand the attacker model previously and derive an analytical model for the attack success probability. Moreover, experiments should demonstrate the message and the time overhead for reaching the consensus. Finally, the authors should compare the proposal with previous work, such as ELASTICO.
After careful consideration, we have re-adjusted the content of the article and summarized your recommendations into three aspects. Here are our responses:
Aspect 1: Attack models
Response 1: Thank you for your suggestion! We have reorganized the structure of section 4.1. This section first introduces three attack models include 1% attack, corruption attack and Byzantine node attack, then, we verify theoretically and experimentally that the proposed algorithm can better cope with these three types of attack models.
Aspect 2: Lamport's liveness and safety properties
Response 2: Thank you for your suggestion! By re-reading our article we found that our analysis of Lamport's liveness and safety properties is not quite right yet. So, we combine the analysis of Lamport's liveness and safety properties with the analysis of the attacker model. Details are written in section 4.1.
Aspect 3: Motivation mechanism
Response 3: Thank you for your suggestion! We have conducted further information search on the motivation mechanism in the sharding consensus algorithm and found that there are few existing studies on the motivation mechanism. However, it is necessary to set a motivation mechanism standard. So, we give some tips on how to set up a motivation mechanism in sharding consensus algorithm instead of giving details. Details are written in section 3.4
Aspect 4: Experiments
Response 4: Thank you for your suggestion! We originally wrote this article mainly because most of the existing slicing algorithms still use the traditional 1/3 fault tolerance when using the Byzantine consensus algorithm. In fact, in order to meet the 99% verification success rate, the actual fault tolerance in each slice should be much less than 1/3. So, we only focused on this and neglected to discuss and compare security and complexity. Upon receiving your suggestion, we have realized that it is necessary to analysis the effectiveness of the algorithm, and we have compared the algorithm with the ELASTICO algorithm from three aspects in section 4.3.
Please review our revised draft again, and we hope to receive your approval.
Best regards
Author Response File: Author Response.pdf
Reviewer 3 Report
The paper is clearly written and encompasses the targeted research area with corresponding references. The method is clearly explained and backed up with experimental results. There are just few minor issues that should be resolved:
1. Fig. 1 does not seem to be necessary.
2. Fig. 2 should be increased.
3. Check the paper for spelling errors, e.g. "In this saction", "shade", etc.
4. Performance analysis in Table 1 is not clear. Please use more precise way (e.g. numerical values) to indicate the performance levels, and add corresponding references for these data.
5. In listing possible applications of blockchain (Introduction section), consider to include also digital transformation of metrology, e.g.
https://www.mdpi.com/1424-8220/22/13/4708
6. Please add content about possible application of your algorithm, i.e. are there some special niches where your algorithm is better suited than the alternative ones (and vice versa)?
Author Response
Dear Reviewer:
Thank you for your letter and your suggestions on our manuscript. These comments are invaluable and help better notice the details in our articles and provide important guidance for subsequent research.
Here are our responses:
Point 1: Fig. 1 does not seem to be necessary.
Response 1: Thanks for reviewer’s comment. After careful consideration, we find that this figure has been included in the reference 7. Therefore, we have removed Figure 1 based on your comments
Point 2: Fig. 2 should be increased.
Response 2: Thank you for your suggestion! We are very sorry for our careless mistake that the words in Fig. 2 are very blurry, and we have re-edited and enlarged this figure.
Point 3: Check the paper for spelling errors, e.g. "In this saction", "shade", etc.
Response 3: Thank you for your careful review of the article! We are very sorry for our careless mistake. We have checked this article again and corrected the spelling mistakes.
Point 4: Performance analysis in Table 1 is not clear. Please use more precise way (e.g. numerical values) to indicate the performance levels, and add corresponding references for these data.
Response 4: Thank you for your suggestion! We have replaced “Complexity” column with numerical values to indicate the distinctions more clearly, and the corresponding references for these data are in reference 7
Point 5: In listing possible applications of blockchain (Introduction section), consider to include also digital transformation of methodology, e.g.
Response 5: Thank you for your suggestion! We have compared these two articles and found that the article you recommend is more suitable and superior than the original one.
Point 6: Please add content about possible application of your algorithm, i.e. are there some special niches where your algorithm is better suited than the alternative ones (and vice versa)?
Response 6: Thank you for your suggestion! We have further evaluated our algorithm and found three advantages and some possible applications which are listed at the end of Section 4.
Please review our revised draft again, and we hope to receive your approval.
Best regards
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
Although the authors have made an effort to provide more detail on the attack model, the paper still lacks formalism in proving the properties of the consensus and that the proposal is resilient against the listed attack models. Furthermore, the authors confuse the attacker model with the experiments' description and the security analysis of the proposal. Therefore, it is suggested that the authors separate, into distinct sections, the generic attack models, substantiating the importance of resisting these attacks, from the security analysis that reinforces how the proposal resists attacks.
Author Response
An Effective Sharding Consensus Algorithm for Blockchain System
(Manuscript ID: electronics-1793548)
Response to Reviewer 2 Comments
Dear Reviewer:
Thank you again for your letter and your suggestions on our manuscript. These comments are invaluable and help we recheck our article and point us in the direction of improvement.
Here are the points you listed:
Point: Although the authors have made an effort to provide more detail on the attack model, the paper still lacks formalism in proving the properties of the consensus and that the proposal is resilient against the listed attack models. Furthermore, the authors confuse the attacker model with the experiments' description and the security analysis of the proposal.
Your suggestion: Therefore, it is suggested that the authors separate, into distinct sections, the generic attack models, substantiating the importance of resisting these attacks, from the security analysis that reinforces how the proposal resists attacks.
Based on your suggestion, we recheck the previously revised section 4.1 and find that our analysis of the security of the algorithm only stops at using a certain method to solve a certain problem and does not carefully analyze the details of the specific algorithm. Meanwhile, we also notice that although the entire section 4.1 is described separately according to the different attack models, the overall content is rather messy.
After careful consideration, we have rewritten section 4.1 based on your suggestions.
Here are our responses:
Response: Thank you for your suggestion! We have reorganized the structure of section 4.1. We first briefly introduce the three attacker models considered in this paper. Then, we conduct a detailed and specific security analysis based on these three models and the specific details of the method proposed in this paper. Each part targets a particular attack model. Meanwhile, in order to verify the effectiveness of the proposed method when analyzing the Byzantine node attack problem, we conduct additional experiments. We also make supplementary analysis and explanations for the special circumstances that occur in the experiment so that the entire Section 4.1 is more complete.
Please review our revised draft again, and we hope to receive your approval.
Best regards
Author Response File: Author Response.pdf
