Proof of Pseudonym: Blockchain Based Privacy Preserving Protocol for Intelligent Transport System

Intelligent Transportation Systems is the future for safe and secure transportation. Vehicles in the ITS share basic safety information which can prompt the disclosure of the real identity of the vehicles. Thus, adversaries can misuse these safety messages. Pseudonyms are alias granted to vehicles by trusted authorities to conceal their original identities. To avoid linkability, various pseudonym generation and distribution protocols have been proposed. Such protocols pose overheads in the system as they are performed by Central Authorities. Therefore, re-utilizing the existing pseudonyms through shuffling is the most optimal mechanism for ITS. The Blockchain is a digital ledger and tamper-resistant record of transactions. It eliminates the need of central authority as well as provides anonymity of transactions resulting in more secure and privacy protected solution. To handle distribution optimization issue in the pseudonym shuffling process without a central authority, the blockchain is used with its distributed consensus. The shuffling results are logged in blocks as transactions. Pseudonym shuffle randomness is achieved via blockchain and it provides robustness in the structure. When one system fails, the rest would continue to work. The method also provide fully traceable record in case of certification revocation. The existing blockchain-based pseudonym shuffling mechanism uses traditional consensus algorithms to support the cryptography operation. This leads to overhead in terms of execution time and memory usage. This research proposes Proof of Pseudonym consensus protocol for the shuffling scheme to improve the efficiency of consensus as compared to Proof of Work, Proof of Kernel Work and, Proof of Elapsed Time in terms of time and memory. The execution time of Proof of Pseudonym is shorter than other algorithms. The security and privacy analysis revealed that our scheme achieves identity privacy, unlinkability, and non-repudiation properties. Threat analysis evaluates the proposed protocol in terms of both internal and external attacks.

consensus is suggested for pseudonym management over 65 blockchain but they need to be tested and compared.

66
A. RESEARCH CONTRIBUTIONS 67 The main contributions of the paper are as follows.

84
The rest of the paper is divided as follows. Section II 85 discusses the related work. In section III we discuss the con-  Pseudonyms can be used to protect the vehicle from mali-97 cious attacks [14]. It is suggested by D. Forster   The second blockchain is supervised by the vehicles with 70 read-only rights and the RSUs (having write access rights). technology. The shortcoming of this paper is the complexity 80 of implementation. The problem defined by K. Shi et al. [28] 81 is that both the vehicles' and users' privacy can be disclosed,  i.e. location privacy. Pseudonym management is done by used to share data without relying on RSUs. The consensus 1 used is delegated proof-of-stake algorithm which is prone to 2 attacks discussed above. The scheme proposed by J. Ma et al. 3 [30] is a blockchain scheme using RSU for an attribute-based 4 encryption algorithm. In their scheme the target is to protect 5 open cloud access environment from unauthorized access.

6
A privacy-preserving pseudonym management framework 7 proposed by S. Bao et al. [31] which is more cost-effective 8 than the existing approaches. There are mainly two contribu-9 tions in the paper, first is the use of blockchain technology 10 for pseudonym management and the second is the VCS 11 pseudonym certificate shuffle scheme. It reduces the cost of 12 pseudonym management and generation. Privacy Manager 13 (PM) which is distributed is also introduced which aims to 14 improve the network robustness and to ease the computation 15 burden on RSUs. If a malicious user compromises a PM or if 16 a PM loses its link with a blockchain, the entire blockchain 17 will remove the PM after repeated failed attempts to obtain 18 its response. The PKI will abandon all the pseudonym sets 19 and they will not be used any further. Total processing time 20 varies from 0.2 seconds for 100 transactions to 2 seconds for 21 1,000 transactions. The problems identified are: The shuffle 22 management in a cloud is not explained clearly i.e. which   There is no answer or solution provided for the case as     48 Blockchain is a decentralized distributed network that pro-    We set the guessing characters for both the algorithms in 99 order to provide the search set for the puzzle to find the string 100 (puzzle) in the given set.     a. Nn: Total number of client nodes b. Nc: Client node (Available objects) 2: Begin Procedure 3:
mous IDs in a domain.   Table 1.

74
• PKI will be used for the initial registration of vehicles 75 i.e it generates and broadcasts the permanent id, pk, sk 76 and cert to manufacturers via a secure channel.

77
• Manufacturer will give the above credentials to vehicles.   Step 1: PKI generate constant pid i , cert i , pk i , sk i 3: Step 2: PKI send pid i , cert i , pk i , sk i through secured channel to Manufacturer. 4: Step 3: Manufacturer issue PM i collects all the used (expired) pseudonyms i.e., PID ⊇ ∪ ∞ i=0 PID i from RSU i . 13: Counts the number of used PID = n; 14: Encapsulates PID into package and sends to PM cloud network.

15:
PM i picks PID sets in PM cloud.

16:
PM i shuffles the PID sets in PM cloud and relocate them to destination PM i . 17: end for 18: PM i starts mining on the selected PID sets. 19: The block is broadcasted into the network by winner miner. 20: for {y=1 y i; y++ } do   RSU i records the PID as tx in BC and starts mining.

6:
After the block is published RSU i distributes the PID to V n .

7:
V n returns PID to RSU i after using.

8:
RSU i starts mining again for used PID i 9: RSU i sends the used PID i to PM i 10: end for 11: Exit E. PROOF OF PSEUDONYM 13 Proof of Pseudonym working as client and cloud server is as 14 follows.      However, if C I P is selected for mining and time is 62 generated for selected nodes, step 1 is performed oth-63 erwise step 4 is carried out. We also calculated the 64 best, average, and worst time complexities of the three 65 algorithms in Table 2. that a simple CPU is not suitable for solving a difficult puzzle. 26 We also constructed a PoW 2 with difficulty (leading zeros) 27 and a hash consisting of data, timestamp, and nonce where 28 the guessing is based on the hash. We also show in Figure 8 29 the average pageable memory taken by the puzzles (nonces) 30 in kilobytes in CPU. The graph shows that as the puzzle 31 difficulty level increases, the memory is more occupied.    with PoW1 and PoW2 can be seen in Figure 13. In Figure   11 14 Proof of Pseudonym is compared with time cost of trans-12 actions calculated using PoW by [31]. We can see in Figure   13 14 that  28 We show how this study overcomes the attacks discussed 29 above. The IBA cannot get any useful information from its 57 vehicles or even if it tries to obtain the pseudonym sets.

58
RSUs will identify the malicious activity as it records 59 the transaction over a blockchain.

60
However, the ITA will try to use the pseudonyms  for mining. In case PM is compromised, the CA will 89 discard it and the network will discard its pseudonyms.  The authors would like to acknowledge the support of Prince blockchain based certificate revocation scheme for vehicular communica-