PAS: Privacy-Preserving Authentication Scheme Based on SDN for VANETs

Round 1

Reviewer 1 Report

In this paper the authors propose a privacy-preserving VANET authentication scheme built on software defined networks infrastructure. The presented solution to the problem is interesting and seems to be on the right direction.
The paper seems to be well organized in exception of some points that need addressing.

Major Comments:
1. Introduction:
The introduction does not provide sufficient background information. The introduction starts directly with the problem context without providing some introductory background to familiarize the reader with the problem environment. I suggest adding a brief introduction to VANETs and elaborate more on why they're a special case of MANETs.

2. Related Works:
I'd recommend extending this section to cover more state of the art. As you described at the beginning of the paper, this is a rapidly growing research area. Another research direction was focused on the use of blockchain in authentication, which would be something useful to compare to.

3. Preliminaries:
I suggest adding a subsection here addressing 

4. Proposed scheme, security, and performance analysis:
The threat modelling part is very clear. Actually, all of section 4 is clear and easy to follow.
The proposed authentication mechanism has been proposed earlier and was proved to be secure in references [28] and [30], as explained in section 5. This makes me wonder how this authentication contributes ot the paper's contribution. Perhaps, adding further description that would distinguish how the proposed method is different from what was done before, especially with relevance to the privacy preservation.

5. The paper is missing a discussion section where the obtained results and the proposed system is compared to other systems. It is difficult to establish the current system's superiority based on the simple table mentioned in the related works section.

Minor comments:
1. In table 1, mention the author’s name with the reference number.
2. The language of the paper needs some grammatical review to make it easier to understand.

Author Response

Thank you for your positive comments and valuable suggestions to improve the quality of our manuscript. We have revised our manuscript regarding several issues mentioned above, as follows.

1. We supplemented the background information and descripted the advantage of VANETs to familiarize the readers with the problem environment.
2. We added subsection 2.5 blockchain-based authentication schemes and analyzed latest authentication schemes based on blockchain (BUA, BPPA and EPAM). Meanwhile, the summary and comparison of blockchain-based authentication scheme are added in table 1.
3. We supplemented subsection 3.1 VANETs, which includes the VANETs architecture and the details of two different communication types.
4. We revised the content in subsection 5.1-ii Authentication and added further description to show the contribution of PAS.
5. We added section 6 Discussion to conclude the advantage of PAS compared with the tradition schemes. Meanwhile, we briefly discussed several open issues.
6. Finally, we added authors’ names in table 1 and corrected syntax errors.

Reviewer 2 Report

First of all, good work.

I found it interesting, however I have a few comments to make.

• Table 1, please add the scheme names and not only the cite ref. Will be easy to read in that way.
• Table 1, your own solution (PAS) should be included there.
• The DMV entity, in a more general IoT scenario, could be the manufacturer itself? Also, can the TA be integrated in the DMV?
• Can the TA be a distributed entity?
• What happens if TA is compromised, or lose connectivity ?
• How is the trust model for DVM and TA? I've readed that we assume that are fully trusted entities, but I consider that they are critical elements of your solution and for that, I was expecting some clarification. Lines 233-235.
• The network architecture does not provide IdM description but appears in the figure.
• Figure 3 shows the step: "Add vehicle to blockchain". Are we using a blockchain too? I've not seen further details about that.
• Section 4.4, regarding the vehicle registration needs a sequence diagram like others.
• Line 323, I'm not sure what do you mean with "if a secure bi-tunnel needs to be established"
• Figure 4. Seems to have some typos regarding the challenges. I'm not 100% sure, but please check it.
• How is the missbehaviour detection carried out? Another vehicle can conmunicate it to the infrastructure but, can the infrastructure itself detect it? Do you have any monitoring element?
• How is data shared across RSU domains? Figure 6, step 6.
• The credentials are "only" pseudonyms with limited information? Can they have other attributes? I think that you should put more details about how the credentials are.
• You should include tech details about anonymity/privacy preserving security considerations to suppor the conclusion. I mean, I see some assumptions regarding that, but seems to be necessary to detail that assumptions.

Author Response

We would like to sincerely thank you for your advices and constructive comments. We have revised our manuscript regarding several issues mentioned above, as follows.

1. We added authors’ name in table 1 to make it easier for readers to understand the content in the table 1.
2. We supplemented PAS in table 1 to facilitate readers to understand the advantages of PAS more intuitively.
3. In VANETs architecture, DMV provides offline services for vehicles (e.g., apply for a real ID, renew a driver's license) and TA provides online services for vehicles (e.g., vehicle registration based on real ID, public/private keys issuance). In the most schemes, DMV and TA are separated, or focus on TA and ignore the details of the real ID issued by DMV to make it easier for readers to understand the function of each entity in VANETs.
4. In the authentication schemes based on DSRC/WAVE standard, TA is set as a single entity. Recently, the blockchain based authentication schemes attempt to set TA as a distributed entity to resist the risk of TA being compromised. However, the high computational cost and communication cost caused by consensus mechanism bring more additional burden to TA. As a result, before the computational and communication overhead of blockchain decreases significantly, we believe that it is not worthwhile to separate TA as a distributed entity for VANETs with high requirements for efficiency at present.
5. If TA lose connectivity, unregistered vehicles are not able to join VANETs through registration with TA and illegal vehicles cannot be removed from VANETs through revocation protocol, which will threaten the security of VANETs. If being compromised, TA cannot obtain the location obtain vehicle location privacy or revoke any legal vehicle since PAS meet the privacy requirements of distributed resolution authority. However, in DSRC/WAVE standard and most authentication schemes, TA is maintained by the government or regulator. Therefore, we believe that it is reasonable to regard TA as a completely trusted entity.
6. DMV supports offline services for vehicle. The owner of vehicle needs to go to the DMV to register the vehicle offline, so DMV is safe by default. As TA is maintained by the government or regulator, it is reasonable to assume that TA and DMV can communicate securely and reliably.
7. We supplemented the function of SDN RSU controllers and kept the consistency of content and figure.
8. Owing to our negligence, we are regret to bring you trouble. We tried to use the blockchain to achieve VANET authentication. However, in the simulation experiment, the high computational cost and communication cost of the blockchain made the base station unable to respond to the RSU update data request in time, which affected the vehicle authentication efficiency. Therefore, we believe that it is inappropriate for blockchain to be applied to VANETs authentication schemes until it is rapidly improved.
9. We revised the content of Figure 4 and supplemented the details of vehicle registration protocol, which guarantees the consistency of Figure 4 and other sequence diagrams.
10. We revised the description to make it easier for readers to understand what we want to elaborate.
11. We have checked and corrected the typos in Figure 4 (As shown in Figure 5).
12. For example, when a vehicle sends malicious traffic information to surrounding vehicles through BSM message (BSM contains traffic information, pseudonym and signature of the vehicle), the surrounding vehicles forward the BSM sent by the malicious vehicle to RSU. RSU are able to confirm the content of the message and the identity information of the signer. The malicious vehicle cannot deny the message sent. Then, vehicle revocation protocol is executed and remove the malicious vehicle in time.
13. The proposed scheme adopts SDN data sharing mechanism to achieve data shared across RSU domains.
14. In the traditional scheme, the credentials include the vehicle identifier, expiry time, public key, and issuer’s signature. Since expiry time and public key are maintained in SDN RSU controller and issuer’s signature is removed. It is enough to only store the vehicle identifier information in the credentials. Since PAS does not consider attribute-based services, it is sufficient for the vehicle to only provide vehicle identifier.
15. We revised the content in subsection 5.1-ii Authentication and added further description to show how PAS resist the attacks from external adversaries.

Reviewer 3 Report

The authors should make sure that all citations and references are in the correct sequence.

• Explain the nature and family of hash and encryption functions that has been considered in the proposed scheme?

• To understand the network model, the authors need to provide details about the types of communication links between the used devices.

• Authors should thoroughly check all the symbols used in the proposed scheme.

• The conclusion part should provide future directions, which will assist other scholars in conducting study in this field.

• Finally, authors need to extensively proofread the article with the use of English.

Author Response

We feel great thanks for your professional review work on our paper. According to your nice suggestions, we have made corrections to our previous draft, the detailed corrections are listed below.

1. We have checked and corrected all citations and references sequence.
2. We supplemented the description of the algorithms used in the proposed scheme, and added the cipher library used in Table 6.
3. In PAS, the communication links type between vehicle and RSU in authentication is based on DSRC/WAVE standard. We have added subsection 3.1 to make it easy for readers to understand the communication details of the proposed scheme.
4. We have checked and corrected all the symbols used in PAS.
5. We added section 6 Discussion to highlight the advantage of PAS compared with the tradition schemes. Meanwhile, we briefly discussed several open issues to assist other scholars in conducting study in this field.
6. We checked the grammar carefully and corrected some syntax errors.

Round 2

Reviewer 1 Report

Thanks are due to the authors for addressing the comment made by the reviewers.

I have no further comments.

Reviewer 2 Report

Dear authors, thank you very much for your corrections and clarifications. In this version the result is much more consistent and I have no further comments.