Highlights

A Survey on Ethereum Pseudonymity: Techniques, Challenges, and Future Directions

Shivani Jamwal,José Cano,Gyu Myoung Lee,Nguyen H. Tran,Nguyen Truong

• This insightful survey provides a novel systematic analysis of privacy preservation techniques in conjunction with
the state-of-the-art Deanonymisation methods in Blockchain and Ethereum ecosystems following novel systematic
scrutiny.

• The paper also suggests applicable and potential techniques for the countermeasure of those Deanonymisation methods,
exploring the intricate dynamics of privacy and deanonymization in the Ethereum blockchain ecosystem.
• It also examines the regulatory-compliance challenges and their implications arising from the intersection of the GDPR
and Ethereum, which is vital for ensuring the coexistence and advancement of groundbreaking blockchain capabilities
and protecting user data.
A Survey on Ethereum Pseudonymity: Techniques, Challenges, and
Future Directions
Shivani Jamwala,∗ , José Canoa , Gyu Myoung Leeb , Nguyen H. Tranc and Nguyen Truonga,∗∗
a School of Computing Science, University of Glasgow, Glasgow G12 8RZ, United Kingdom
b School of Computing Science, Liverpool John Moores University, Liverpool L3 3AF, United Kingdom
c School of Computer Science, The University of Sydney, Darlington, 2008, NSW, Australia.

ARTICLE INFO Abstract

Keywords: Blockchain technology has emerged as a transformative force across various sectors, including
Anonymity finance, healthcare, supply chains, and intellectual property management. Beyond Bitcoin’s role as
Bitcoin a decentralized payment system, Ethereum represents a notable application of blockchain, featuring
Blockchain Smart Contract functionality that enables the development and execution of Decentralised Appli-
Data Protection Regulation cations (DApps). A key feature of Ethereum, and public blockchains in general, is pseudonymity,
Deanonymization typically achieved by using public keys as pseudonyms for users. Despite implementing several
Ethereum privacy-preserving techniques, the public recording of users’ activities on the blockchain allows
GDPR various deanonymization methods that can profile users, reveal sensitive information, and potentially
Privacy Preservation re-identify them. Most blockchains, such as Bitcoin, Litecoin, and Cardano, employ the Unspent
Transaction Output (UTXO) model for accounting, which focuses on individual transactions and
is susceptible to various deanonymization techniques. Conversely, Ethereum uses an Account-based
transaction model, integrating the concepts of accounts and wallets at the protocol level. This makes
most UTXO-based deanonymization techniques ineffective for Ethereum. Nonetheless, alternative
methods have been proposed and developed with the potential to deanonymize Ethereum users.
Privacy preservation techniques have been employed to counter deanonymization attempts; however,
the challenges related to these techniques, their effectiveness and efficiency, and the trade-off between
usability and protection levels remain areas for further exploration. This survey presents a compre-
hensive analysis of state-of-the-art privacy preservation along with deanonymization techniques in
Blockchain and Ethereum ecosystems. It examines the intrinsic mechanisms supporting pseudonymity
in Ethereum, providing a detailed assessment of the advantages and disadvantages of privacy
preservation techniques, and suggesting potential countermeasures against those deanonymization
methods. It also discusses the implications arising from the intersection of DApps and data protection
legislation, which is vital for ensuring the coexistence and advancement of groundbreaking blockchain
capabilities and protecting user data.

1. Introduction critical elements that define Ethereum, and more broadly,

the field of cryptocurrencies, is its pseudonymity, which,
In recent years, blockchain technology has emerged as a while providing users with a level of privacy, also makes it
transformative tool with significant implications for various a hub for illicit activities. Therefore, the deanonymization of
sectors, from finance to healthcare and supply chain manage- Ethereum transactions, which refers to identifying the par-
ment to digital identity verification. A prominent manifesta- ties involved, has become an increasingly important research
tion of blockchain technology is Ethereum, a decentralized,
domain [18].
open-source platform featuring smart contract functionality
The pseudonymous nature of any blockchain system,
that enables the development and execution of decentralized Ethereum in particular, is achieved using public keys as
applications (DApps). Since its inception in 2015, Ethereum pseudonyms for users. However, users’ activities can be
has transformed the landscape of the traditional client-server traced and linked by analyzing these public keys and trans-
applications and services (e.g., DApps), the financial world action patterns [200]. Several techniques and tools have
(e.g., cryptocurrencies and Decentralized Finance DeFi in
been developed over the years that have the potential to
general), and the digital marketplace (e.g., non-fungible
deanonymize these transactions, resulting in the ability to
token NFT and decentralized storage services) [32]. With profile and re-identify the users [42, 18, 77]. To our knowl-
a robust programming language (i.e., Solidity) and an open- edge, most deanonymization techniques have targeted Bit-
source, decentralised platform, Ethereum has gained signif- coin and other blockchains using the Unspent Transaction
icant attention worldwide, and it is now the second-largest
Output (UTXO) accounting model due to the feasibility
cryptocurrency network after Bitcoin. However, one of the
of inferring sensitive information about on-chain activity
∗ First
Author, Corresponding Author by examining and tracking UTXOs. In the account-based
∗∗ PrincipalCorresponding Author model of Ethereum, an Ether token (i.e., coin) belongs to
[email protected] (S. Jamwal); an account, which is specified by public/private key pairs,
[email protected] (J. Cano); [email protected] (G.M. Lee);
[email protected] (N.H. Tran); [email protected] (N.
represented as its balance [32]; there are no such kinds
Truong) of reference linking tokens in transactions in the UTXO
ORCID (s):

Jamwal et al.: Preprint submitted to Elsevier Page 1 of 26

 Survey on Ethereum Pseudonymity

model [134]. Consequently, those deanonymization tech- Table 1

niques are impractical for Ethereum and other account-based Abbreviations used in the paper.
blockchain platforms because of the elemental difference in ABE Attribute-based Encryption
their transaction model. However, the account model has to DApps Decentralized Applications
come with the reuse of account addresses. This is generally DPoS Delegated Proof of Stake
a detriment to Ethereum users’ privacy because transactions EOA Externally Owned Accounts
linking to an account can effortlessly be traced back to a EEA European Economic Area
single owner. EU European Union
Deanonymization methods that have the potential to EVM Ethereum Virtual Machine
interpret and classify Ethereum transactions have been FHE Fully Homomorphic Encryption
proposed and implemented, resulting in profiling and re- GDPR General Data Protection Regulation
HE Homomorphic Encryption
identifying Ethereum users. At the same time, to counter-
IPFS Inter-Planetary File System
measure those privacy attacks, several privacy-preservation NIZK Non-Interactive Zero-Knowledge Proofs
techniques including Mixing services, Zero-Knowledge- P2P Peer-to-Peer
Proof (ZKP) [34], and Secure Multiparty Computation PbD Privacy by Design
(SMPC) [207] have been developed to empower Ethereum PHE Partially Homomorphic Encryption
privacy [161] [214]. Yet, the challenges associated with PoA Proof of Authority
these techniques, their efficiency, and their scalability still PoH Proof of History
need to be explored. This survey presents an in-depth QAP Quadratic Arithmetic Program
analysis of such state-of-the-art techniques by inspecting the SMPC Secure Multi-Party Computation
intrinsic mechanisms of Ethereum that support anonymity, TGA Transaction Graph Analysis
evaluating various deanonymization techniques, discussing UTXO Unspent Transaction Output
ZKP Zero-Knowledge Proof
potential implications on privacy, and identifying future
zk-SNARK Zero-Knowledge Succinct
research directions in the field of privacy in Ethereum-based Non-Interactive Argument of Knowledge
platforms. zk-STARKs Zero-Knowledge Scalable
This paper presents a survey of the pseudonymity of Transparent Arguments of Knowledge
Blockchain, particularly Ethereum, ecosystem, providing
a detailed study of existing privacy preservation in con-
junction with deanonymization techniques, examining the pseudonymity of Ethereum, including the regulatory con-
challenges researchers and practitioners face in the line of cerns. Finally, Section 7 concludes the paper with potential
research, and suggesting applicable and potential techniques future research directions.
for the countermeasure of those state-of-the-art Deanonymi-
sation methods. The motivation for this survey is multifold. 2. Background and Challenges
Firstly, it seeks to consolidate the vast amount of research
conducted in the area, catering for a single reference for aca- This section introduces the background of blockchain
demics, researchers, and industry professionals. Secondly, it and Ethereum, including the definition, system model, appli-
aims to highlight the challenges and issues faced with the cation scenarios, and privacy and security in permissionless
state-of-the-art privacy-preservation techniques in permis- blockchain platforms.
sionless blockchain platforms, particularly Ethereum, stim-
ulating further development and improvement. Finally, the 2.1. Blockchain Technology
paper underscores the importance of this research within the Satoshi Nakamoto [134] is credited with the inception of
broader context of data protection legislation, considering Bitcoin and the underlying blockchain technology, a revolu-
the escalating popularity of DApps, ensuring the coexistence tionary decentralized framework designed for secure com-
and advancement of groundbreaking blockchain capabilities putations within networked systems without a central regula-
and protecting user data. tory authority. A blockchain system records all transactions
The remainder of the paper is organized as follows: Sec- or messages transmitted from one network node to an-
tion 2 provides the background of blockchain and Ethereum other. After network nodes’ verification, following a major-
along with the privacy and security, including pseudonymity ity consensus protocol, these transactions are permanently
in these blockchain platforms. Section 3 details the state- recorded on the blockchain, ensuring the immutability of
of-the-art privacy preservation techniques for blockchain stored information and the traceability of each transaction’s
platforms, focusing on strengthening pseudonymity. Section history [188]. From a data management perspective, the
4 discusses the deanonymization techniques by analyzing blockchain serves as a distributed database that archives
their efficiency and potential. Section 5 introduces prospec- an ever-evolving collection of transaction records. This is
tive resolutions for the deanonymization on Ethereum plat- achieved by systematically organizing these records into a
forms and Section 6 ponders the broader implications of the chain of interconnected blocks.
A blockchain system is typically a peer-to-peer (P2P)
network characterized by two types of nodes: miners and
users. The freedom to assume either role lies with each

Jamwal et al.: Preprint submitted to Elsevier Page 2 of 26

 Survey on Ethereum Pseudonymity

node. Miners are pivotal to the blockchain infrastructure,

jointly maintaining the operation of the P2P network [44].
The following discourse provides an in-depth exploration
of the salient components of blockchain technology: the
distributed ledger and the consensus mechanisms.

• Blockchain as a Distributed Ledger: A distributed ledger

is a decentralized database that records all blockchain data Figure 1: Overview of the Blockchain data structure consisting
in a standardized format and is managed by all partici- of blocks linked to each other using cryptographic hashes of
pating miners. It encompasses a series of interconnected the previous blocks.
blocks in a chain linked using a hash function. Each block
is added and maintained chronologically, identifiable by
its unique hash value, i.e., the block address. The block 2.2. Blockchain Characteristics
header incorporates the current version number, the previ- Unique characteristics of blockchain are discussed, delv-
ous block’s hash value, its block address, the Merkle root ing into its core features such as decentralization, autonomy,
hash, and the creation timestamp, as shown in figure 1. transparency, non-repudiation, and immutability.
A consensus mechanism, such as Proof of Work (PoW)
[134], Proof of Stake (PoS) [137], are employed, contain- 2.2.1. Decentralization and autonomy
ing a nonce that confirms the computational correctness of Decentralization and autonomy are quintessential fea-
the block generation. The block body permanently records tures of blockchain technology, breaking away from conven-
all confirmed transactions in the blockchain, which are tional centralized systems.
structured as a Merkle tree, organizing all transactions for The decentralized nature of blockchain technology im-
efficient querying and verification [141]. plies that no single entity controls the entire network [6].
The onus of maintaining the distributed ledger falls on Instead, the network is maintained collectively by all nodes
the miners. They can both access and write data into the participating, often known as ’peers’. Each peer has a copy
ledger. However, before recording any data, its validity of the entire blockchain and participates in the validation of
must be confirmed via the consensus mechanism. While transactions.
users can access data, they can only write and append data This decentralized architecture contributes to the robust-
to the blockchain with miners’ assistance. ness and security of the system. Since no central author-
ity can be compromised, the system is resilient to various
• Consensus Mechanisms: The consensus mechanism is a attacks, including denial-of-service attacks. Moreover, the
fault-tolerant process that allows multiple parties to agree absence of a central authority eliminates the need for trust in
on a single data value or network state [192]. It sustains a single entity, as the integrity of the system is maintained
the blockchain data’s authenticity, consistency, and order by the consensus of the network participants.
across the network. The consensus mechanism resolves The autonomous nature of blockchain technology emerges
the Byzantine general’s problem that in a distributed from its use of smart contracts, which are self-executing
system with numerous nodes, these nodes must reach contracts with the terms of the agreement directly written
an identical decision despite disloyal nodes. Consensus into code. The code and the agreements exist across a dis-
mechanism guarantees the blockchain’s verifiability and tributed, decentralized blockchain network, ensuring their
tamper-resistance by obtaining approval from these nodes execution is automatic, traceable, and irreversible[122]. This
in which mining is the core process to reach consensus on autonomy reduces the possibility of manipulation, bias, or
a newly created block via the blockchain, providing sys- error and enhances efficiency by eliminating the need for
tem liveness and safety. Familiar consensus mechanisms intermediaries.
include PoW [134], PoS [32], Delegated Proof of Stake Decentralization and autonomy, while providing signif-
(DPoS) [184], Proof of Authority (PoA) [50], and Proof icant advantages in terms of security and efficiency, also
of History (PoH) [204]. present challenges. For instance, the autonomous execution
Typically, the security of a blockchain system, like Bit- of smart contracts might lead to unintended consequences
coin, is contingent upon its consensus model. The se- if the code is flawed [164]. These challenges necessitate
curity of consensus is premised on the assumption of ongoing research and development in the field.
an honest majority; that is, the majority of consensus
voting power is honest. Some blockchains, like Bitcoin, 2.2.2. Transparency
incorporate an incentive mechanism to motivate miners Information recorded in a blockchain (i.e., on-chain data)
to create new blocks. This incentive contributes to the is transparent to all nodes in the blockchain network, and
blockchain’s resilience and, based on game theory, aug- users can conveniently access this on-chain data by querying
ments the blockchain’s security and privacy. miners. Transparency enhances data immutability and veri-
fiability because all nodes can detect illegal modification and
data. Nevertheless, privacy leakage due to transparency has

Jamwal et al.: Preprint submitted to Elsevier Page 3 of 26

 Survey on Ethereum Pseudonymity

become a crucial issue that dramatically limits blockchain easily detected. This characteristic benefits security but also
applications [20]. results in the problem that sensitive data cannot be removed
Each transaction on a blockchain is time-stamped, im- from the blockchain.
mutable, and linked to the previous transaction, forming Immutability means whose state cannot be altered after
a chain of blocks. Any alteration in a block impacts all its creation. Immutable transactions make it impossible for
subsequent blocks, making unauthorized changes nearly im- any entity to manipulate and falsify data stored on the
possible without detection [20]. This immutability feature network [138]. Since historical transactions can be audited
enhances transparency and trust in the system, as partici- at any point, immutability enables high data integrity.
pants can verify the integrity of transactions without relying
on third-party intermediaries [215]. 2.3. Ethereum platform
Ethereum, launched in 2015, is an open-source, blockch-
2.2.3. Non-repudiation ain-based platform that enables the development and exe-
Non-repudiation is an inherent property of blockchain cution of smart contracts [32]. The earlier cryptocurrency,
technology that ensures irreversible and undeniable proof Bitcoin, influenced the design of Ethereum, but with signif-
of participation in a transaction or event [66]. The non- icant enhancements to extend its functionalities beyond just
repudiation of Blockchain refers to (i) No one can deny a peer-to-peer electronic cash system.
transaction contents created by himself and (ii) No one can Ethereum, since its inception, has revolutionized the
repudiate the transaction time generated by himself. This blockchain landscape by introducing the concept of pro-
characteristic is vital for creating a trustless environment grammable contracts known as “smart contracts”. This plat-
where participants can confidently interact, knowing their form enables developers to create decentralized applications
actions cannot be denied later. Non-repudiation is achieved that operate on a blockchain, leading to trustless and trans-
in blockchain technology through the use of cryptographic parent interactions over the internet [49].
digital signatures [217]. In a transaction process, the sender In Ethereum, users can create accounts, transfer their
signs the transaction with their private key, which anyone native Ether (ETH) cryptocurrency, and interact with smart
can verify using the sender’s public key. This signature contracts. Smart contracts are immutable and autonomous
serves as strong proof of the origin and integrity of the scripts stored on the Ethereum blockchain that execute
transaction. Once the transaction is included in a block and predefined functions when certain conditions are met [49].
the block is appended to the blockchain, the transaction These smart contracts are pivotal in allowing DApps to
becomes immutable and undeniable [217]. be built on Ethereum, making it more than just a plat-
Due to the characteristic of non-repudiation, if a trans- form for cryptocurrency transactions. One notable fea-
action exists in the Blockchain, it must be initiated by its ture of Ethereum is the introduction of the Ethereum Vir-
signer itself. The node cannot deny that it has published tual Machine (EVM), a runtime environment that executes
the transaction. The distributed nature of blockchain tech- smart contracts. This makes Ethereum a general-purpose
nology ensures that all nodes in the network hold a copy blockchain, unlike Bitcoin, designed with a specific use case
of the blockchain, and this duplication further bolsters non- [49].
repudiation. Suppose a participant attempts to deny their In parallel to Bitcoin, Ethereum offers its users a degree
action. In that case, other participants can refer to their local of anonymity through its use of pseudonymous addresses
blockchain copies to verify the action’s occurrence [217]. linked to each account. These addresses bear no explicit
Non-repudiation, while offering an added layer of se- identifiers tying them to the users’ real-world identities,
curity and trust in blockchain transactions, raises several offering a level of privacy that is essential for many [186].
implications, particularly privacy. Once a transaction is com- Nevertheless, this feature can be a double-edged sword as it
mitted, it is permanently recorded and openly verifiable. opens avenues for potential illicit activities [211].
This feature presents challenges for privacy-concerned users
who may need to obfuscate their activities to maintain their 2.4. Smart Contracts
privacy, bringing forth anonymity and pseudonymity in the A smart contract platform is software that runs on
blockchain. a blockchain, extending its functionality and broadening
its application. Smart contracts are programs stored on a
2.2.4. Verifiability and immutability blockchain that execute when certain predetermined con-
Verifiability means that the validity of each transaction ditions are fulfilled [32]. They are used to automate the
in the blockchain can be verified and cannot be modified execution of an agreement so that all participants can
or removed from the blockchain [138]. Since all miners instantly ascertain the outcome, thus eliminating any need
confirm the blocks in which transactions are recorded via for an intermediary or wasting time. Workflow automation
the consensus mechanism, invalid transactions will not be is possible, triggering the subsequent action when conditions
recorded in the blockchain. Any data modification in the are met [32].
blockchain will be denied unless the adversary compromises Smart contracts are defined in many ways. Szabo first
the whole system. Also, blocks are organized in a chain using proposed that ’Smart contracts are a computable transaction
the hash function, which makes any modification to the data protocol to execute contract terms’ [177]. Ethereum’s smart

Jamwal et al.: Preprint submitted to Elsevier Page 4 of 26

 Survey on Ethereum Pseudonymity

contract is a digital asset control program based on the However, although the security embedded, blockchain is
blockchain [22]. In a narrow sense, a smart contract is a prone to attacks such as double spending, wallet-based at-
program code involving business logic, algorithms, complex tacks (i.e., client-side security), network-based attacks such
relationships among people, legal agreements, and networks. as DDoS, Sybil, and eclipse and mining-based attacks such
A smart contract is a computer protocol that can self-execute as 51% [208], block withholding and bribery [86], [54],
and self-verify after deployment [177]. [172]. As a prominent platform for decentralized applica-
The operation of smart contracts involves three proce- tions, Ethereum presents unique security challenges stem-
dures: contract creation, contract publishing, and contract ming from its open, decentralized, and complex nature.
execution [197]. During contract creation, the contract par- These challenges can broadly be categorised into protocol-
ticipants will negotiate to clarify parties’ rights and obliga- level challenges, smart contract vulnerabilities, and network-
tions, determine the standard contract text, and then program level threats as follows:
them into a smart contract program. The contract program
generally requires auditing for secure execution. In contract • Protocol-Level Challenges: Like all blockchain systems,
publishing, the contract creator signs and requests a miner to Ethereum is subject to attacks that attempt to manipulate
record the signed contract into the blockchain. The contract the consensus protocol. Examples include 51% attacks,
execution is based on an event-triggered mechanism on the where a malicious entity with control over the major-
blockchain, which encompasses transaction processing and ity of the network’s hashing power can manipulate the
preservation mechanisms and is a complete state machine. blockchain, and eclipse attacks, where a node is isolated
Specifically, the external nodes can interact with a smart from the rest of the network and fed false information [65].
contract program by sending transactions. The transactions
• Smart Contract Vulnerabilities: Ethereum’s distinctive
can change the status of the contract. All miners monitor the
feature, the smart contract, introduces unique security
status, and once they detect its change, they execute the smart
issues. These programmable contracts, while powerful,
contract based on its design.
have been a common target for attackers due to vulner-
2.5. Security and Privacy in Ethereum abilities in their code. Notable attacks exploiting smart
contract vulnerabilities include the DAO and Parity wallet
This section delves into the security and privacy con-
incidents. Detecting and eliminating such vulnerabilities
cerns in Ethereum platforms. While the inherent structure of
before contract deployment remains challenging due to
blockchain offers significant security measures, Ethereum,
the Turing completeness of Ethereum’s programming lan-
with its unique characteristics, poses unique challenges in
guage, Solidity, and the immutability of deployed con-
ensuring user data protection and malicious attempts.
tracts [115].
2.5.1. Security • Network-Level Threats: Ethereum is also subject to com-
Security in the blockchain is based on the following mon network-level attacks as part of the Internet. These
factors. First, blockchain technology relies on a decentral- include DDoS attacks, Sybil attacks, and routing attacks.
ized ledger to keep track of all financial transactions. The Effective defence mechanisms are needed to maintain the
“master” ledger would be a point of vulnerability [12]. If robustness and reliability of the Ethereum network [13].
the ledger was compromised, then it could lead to a system
breakdown. Secondly, the ledger exists as a long chain of 2.5.2. Privacy
cryptographically encrypted sequential blocks, reducing the Data privacy of blockchain refers to the property that
risk of data tampering. Blockchain consists of hundreds to blockchain can firstly provide Anonymity where it is the
thousands of unique nodes. Every node has a complete copy state of being anonymous and unidentified [86], [92], [69]
of the digital ledger. The nodes can work independently for and secondly ensure Unlinkbaility where users’ transactions
the verification of a transaction. If the nodes do not agree, related to themselves cannot be linked [169], [162], [69].
then the transaction is cancelled. Despite its innovative applications and significant potential,
Thirdly, The cryptographic keys and two key systems Ethereum grapples with inherent privacy challenges. These
used in blockchain exchanges are very long, complex and issues are related to the system’s pseudonymous nature,
difficult to decipher unless one has the authorisation to view transparency of transactions, and the interaction of smart
the keys. The public key and private key in public-key contracts as follows:
cryptography. Both of these keys are generated using the
Elliptic Curve cryptography method. Firstly, it creates the • Pseudonymity and Linkability: Ethereum accounts are
private key, and then a public key from the private key is pseudonymous, and transactions are publicly visible. This
created using the Elliptic Curve Algorithm (ECDSA) [99]. leaves room for the potential deanonymization of user
Therefore, private and public keys are cryptographically accounts through data analysis. An attacker can link ad-
and mathematically linked. Therefore security in blockchain dresses to identify a unique user or organization and then
ensures confidentiality , availability ,integrity and ledger analyze their financial behaviour. Although this trans-
consistency parency is necessary for ensuring the system’s integrity,
it presents significant privacy concerns [33].

Jamwal et al.: Preprint submitted to Elsevier Page 5 of 26

 Survey on Ethereum Pseudonymity

• Transaction Transparency: Every transaction on the Ether-

eum blockchain is visible to every participant in the
network [142]. While this provides essential advantages,
it can lead to privacy leaks. Information about the value
transferred, the parties involved, and even the timing of
the transaction can reveal sensitive data [130].

• Smart Contract Privacy: Smart contracts, a distinctive

feature of Ethereum, have exacerbated privacy concerns
[108]. The logic and state of a smart contract are visible
to all participants, potentially revealing sensitive business
logic or private data. Moreover, interaction with smart
contracts can leak information about the parties involved.

While these challenges exist, it is essential to note that

continuous efforts are being made by Ethereum developers
and the wider blockchain community to identify and mitigate
Figure 2: Employment of Privacy-Preserving techniques in
these security and privacy threats. Addressing these issues Ethereum considering both computation and implementation
is crucial for adopting and surviving Ethereum as a secure complexities since its inception.
platform for decentralized applications.

3. Privacy Preservation Techniques in data, preserving the privacy of the underlying information
Blockchain Platforms [205]. The Zether protocol brought Differential Privacy to
Ethereum, providing statistical privacy by obscuring indi-
Since the introduction of Blockchain, various privacy- vidual data within aggregated data sets [31]. Collectively,
preservation techniques have continuously been adopted and these advancements have been seen to contribute to the ro-
integrated to enhance the confidentiality and anonymity of bust framework of privacy within the Ethereum ecosystem.
users’ identities and transaction data on the networks. In These privacy-preservation techniques are briefly summa-
this section, we delve into the significant milestones in rized in Table 2 with their pros and cons.
the journey of the privacy evolution in blockchain sys-
tems, particularly Ethereum, from the implementation of 3.1. Mixing Services
Ring Signatures in 2015 to more advancements, such as Mixing services are essential tools in the blockchain
Mixing Services and Differential Privacy. In the journey ecosystem for enhancing user privacy, particularly in ad-
towards enhancing privacy on the Ethereum blockchain, as dressing the non-anonymous nature of cryptocurrencies
seen in Figure 2, several implementations of these privacy- such as Bitcoin. Despite employing pseudonymous ad-
preserving techniques stand out. To name a few, introducing dresses, the public nature of Bitcoin transactions allows for
a unique ring signature scheme using secp256k1 elliptic the analysis and correlation of a user’s transactions. Mixing
curve that provides anonymity for signers within a group services, such as Tornado Cash on the Ethereum platform
[129]and integrating ZKPs to bolster scalability and confi- [147], obfuscate transaction trails to impede address linkage.
dentiality in transactions via zk-SNARKs [131]. Ethereum’s However, they do not protect against coin theft [38]. Notable
privacy mechanisms further expanded with Tornado Cash, a mixing services include MixCoin [29], CoinJoin [52], and
decentralized, non-custodial mixing service that uses smart Private CoinJoin as implemented in DASH [56]. These
contracts and zero-knowledge proofs to enhance transac- services play a pivotal role in preserving user anonymity in
tion privacy [147]. Tornado Cash breaks the on-chain link the blockchain. Several notable mixing services have been
between sender and receiver addresses, providing a higher proposed as follows:
degree of anonymity and privacy for Ethereum transactions
[147]. 1. MixCoin: Introduced by Bonneau [29], MixCoin aims
Additionally, SMPC is used in REN Protocol, an open to provide anonymous transactions for Bitcoin and sim-
protocol built to provide interoperability and liquidity be- ilar cryptocurrencies. It counters passive adversaries by
tween different blockchain platforms [179]. Attribute-based broadening the anonymity set, facilitating simultaneous
Encryption (ABE) technique in Ethereum is the integration coin mixing by all users. Against active adversaries, Mix-
of ABE with smart contracts for fine-grained access control, Coin offers anonymity akin to traditional communication
where data access can be precisely controlled according to a mixes. Significantly, MixCoin incorporates an account-
set of attributes, such as user roles, permissions, or other cri- ability mechanism, deterring coin theft by aligning user
teria, without revealing the identities of the users involved, incentives and fostering rational usage.
particularly relevant for use cases where sensitive data is
involved, such as in healthcare [98]. Lastly, the introduction 2. CoinJoin: Conceived in 2013, CoinJoin offers an alter-
of HE techniques allowed for computations on encrypted native for anonymizing Bitcoin transactions [52, 124].

Jamwal et al.: Preprint submitted to Elsevier Page 6 of 26

 Survey on Ethereum Pseudonymity

Table 2
Comparative Analysis of Privacy-Preservation Techniques in Blockchain Platforms.
Technique Pros Cons
Mixing Services [147, 29, 52, Efficiently obscures transaction trails, making Enhances privacy but do not necessarily in-
124, 56, 9, 24, 168, 199, 201] linking addresses and tracing transactions crease security against theft or loss of cryp-
back to users difficult. tocurrency.
Ring Signature [15, 76, 174, Ensures anonymity of the signer within Do not conceal the transaction amounts or
174, 118, 3, 209, 185, 47, 140, a group, thereby effectively shielding the the recipient’s address. The presence of com-
43] sender’s identity in a transaction. putational overhead associated with larger
ring sizes can affect transaction processing
times.
ABE [202, 198, 98, 28, 37, Supports both confidentiality and access con- Limited adoption of certain DApps is at-
123, 132, 87, 153, 95, 203, trol and has the potential for decentralized tributed mainly to their complexity and the
206, 143] multi-authority systems. challenges associated with implementation.
SMPC [116, 128, 213, 216, Enables confidential computations involving Computationally complex and can lead to in-
17, 5, 17, 220] multiple entities while ensuring the privacy creased network latency, affecting blockchain
of data segments among participants. networks’ efficiency.
ZKP [146, 30, 83, 146, 149, Allows for validating blockchain transactions Introduces significant computational over-
85, 26, 93, 131, 144] in DApps without revealing sensitive data and head, especially in non-optimized implemen-
offers strong cryptographic privacy guaran- tations.
tees.
Differential Privacy [8, 59, 60, Offers strong mathematical assurances May lead to diminished data utility, impact-
1, 105, 78, 125, 90, 73, 114, against data breaches, crucial in protecting ing the effectiveness of data-driven deci-
97, 96, 145] individual privacy during computations sions in blockchain applications. Privacy level
pertinent in DApps for processing or sharing depends on parameter configuration, which
sensitive data. requires careful balancing to achieve desired
privacy without significantly compromising
data usefulness.
HE [79, 53, 133, 189, 113, Enables computation on encrypted data while Complex implementation and computation-
189, 167] preserving privacy; ideal for DApps which ally intensive which results in slower perfor-
manage and process sensitive data either on- mance, less scalable, and impracticality for
chain or off-chain. smart contract operations.

Rooted in the concept of joint payments, CoinJoin allows Mixing services represent a pivotal advancement in pre-
a user to collaborate with another user to execute a serving user anonymity within the blockchain realm, espe-
collective payment within a single transaction. Such joint cially given the transparency challenges posed by cryptocur-
payments considerably diminish the likelihood of tracing rencies like Bitcoin. While they mask transactional histories,
input-output links or discerning a specific user’s mon- inherent vulnerabilities remain, particularly regarding coin
etary flow direction. Early mixing services employing security. As blockchain technology progresses, a deep un-
this approach, like SharedCoin [9], relied on centralized derstanding of these services’ pros and cons becomes crucial
servers. However, such centralization, while simplifying for developers and users [168].
the process, introduced trust issues. Users had to trust
these service operators to safeguard the bitcoins and 3.2. Ring signature
not retain transaction logs, which would undermine the Ring signatures, introduced by Rivest, Shamir, and Tau-
privacy efforts. man [156], represent a sophisticated approach to digital
signatures that allow a member of a group to sign messages
3. Private CoinJoin: DASH has implemented an advanced anonymously. An example of ring signatures in practical
version of CoinJoin [56] providing privacy features throu- use is in anonymous voting systems [15], where they ensure
gh its 𝑃 𝑟𝑖𝑣𝑎𝑡𝑒𝑆𝑒𝑛𝑑 function [56, 24, 52]. 𝑃 𝑟𝑖𝑣𝑎𝑡𝑒𝑆𝑒𝑛𝑑 voter anonymity while maintaining vote integrity.As Figure
is an optional feature in DASH that blends multiple 3 illustrates, the ring consists of several members, any of
transactions, making it substantially challenging to deter- whom could be the actual signer, depicted by the figures
mine the source, destination, and amounts in individual around the “RING”. This ensures that while the signature
transactions. This CoinJoin-inspired method allows users validates the message as originating from the group, the
to utilise the increased privacy feature without making it actual author’s identity remains concealed, depicted by the
mandatory for all DASH transactions [52]. detached figure holding the “Signature Value”.

Jamwal et al.: Preprint submitted to Elsevier Page 7 of 26

 Survey on Ethereum Pseudonymity

Such a mechanism of signer ambiguity, as shown in the

diagram, ensures privacy and is computationally formidable
to penetrate, making it a valuable tool for maintaining the
signer’s anonymity within a set [76, 174]. The practicality
of ring signatures is evident in various domains, most no-
tably within blockchain technology and cryptocurrencies,
where they play a pivotal role in safeguarding transactional
anonymity [174, 118].
In the realm of cryptocurrencies, ring signatures find a
significant application in Monero[185], a prominent privacy-
centric cryptocurrency. By leveraging the ring signature
mechanism, Monero can obscure the actual sender of a
transaction by blending their identity with decoy inputs, Figure 3: Privacy Preservation techniques based on Ring
ensuring the origin of funds remains confidential. Techni- Signature where the level of anonymity is restricted to the
cally, Mereno combines Ring Signatures, Stealth Addresses number of users (i.e., size of the ring).
[47], and Confidential Transactions [140] to achieve trans-
actional privacy. Ring Signatures, in particular, involve the
creation of a signature by a group such that it becomes users cannot access data beyond what their private keys
computationally infeasible to determine the actual signer. allow, even in instances of collusion [37, 123].
Stealth Addresses ensure transaction outputs are directed A unique feature of ABE is its capacity to support con-
to one-time addresses, rendering them unlinkable to the fidentiality and access control through singular encryption
recipient’s public address. This prevents potential linkage [84, 190, 212]. This is facilitated by the involvement of
attacks where transaction patterns are analyzed to determine four key parties, namely, the cluster head or data owners,
sender-receiver relations. These techniques combined make blockchain miners, attribute authorities (AA), and the dis-
Monero transactions opaque, ensuring the details of the tributed ledger (typically represented by blockchain with
sender, receiver, and transaction amounts remain concealed blocks of transactions) [132, 87, 153]. Despite its potential,
[185]. the application of ABE remains limited, primarily due to a
However, while ring signatures offer enhanced privacy, general lack of comprehension surrounding its core princi-
it is imperative to understand their limitations and the trade- ples and efficient implementation strategies [95, 203].
offs involved [43]. Specifically, the size of the ring (i.e., Furthermore, ABE does not necessitate a fixed author-
the number of potential signers) can impact both the level ity, enabling the potential for multiple authorities within
of anonymity provided and the computational overhead re- a decentralized network, providing a robust and adaptable
quired. Larger rings offer higher anonymity but at the cost approach for network management [36, 206]. Technologies
of increased computational resources and transaction sizes. such as the Inter-Planetary File System (IPFS) [16] and
Moreover, while ring signatures shield the sender’s identity, Storj [195] leverage this, permitting witnesses to assume the
they do not inherently conceal transaction amounts or the role of these authorities within a blockchain. Despite the
recipient’s address. To achieve a more comprehensive pri- advances, implementing ABE within a blockchain-centric
vacy solution, ring signatures are often combined with other approach presents ongoing challenges. It is a topic of active
cryptographic techniques, such as Confidential Transactions research [143].
[140] or Stealth Addresses [47] as pointed out in Monero.
This combination will complement Ring Signatures as a 3.4. Secure multi-party computation
robust cryptographic tool, balancing transactional privacy Secure multi-party computation (SMPC) has emerged
and computational efficiency. as a cornerstone cryptographic protocol, enabling multiple
entities to collaboratively compute a function over their
3.3. Attribute-based encryption respective inputs while maintaining the confidentiality of
Attribute-based Encryption (ABE) is an advanced cryp- these inputs [55, 48]. Rooted in the seminal work by Yao in
tographic method where user attributes are defining and 1982 [207], SMPC distributes data or programmatic states,
regulatory factors in the encryption process that provide the such as those inherent to blockchain-based smart contracts,
framework for the ciphertext produced by a user’s secret across 𝑁 parties via advanced secret sharing techniques.
vital [160, 21, 36]. A practical application can be seen in Under this paradigm, a subset of 𝑀 parties from the total 𝑁.
healthcare blockchain platforms, where ABE controls access 𝑁 is indispensable for the joint computation of a designated
to sensitive patient data [98]. Successful decryption of the input, producing the output and revealing the underlying data
encrypted data using a user’s secret key occurs only if the or programmatic states. Notably, each participating entity
user’s attributes align with the ciphertext attributes [28]. is privy only to an input segment, thereby possessing a
This arrangement highlights the crucial security aspect of unique position on a distinct polynomial, instrumental for
ABE known as collision resistance, ensuring that malicious discerning a specific data segment [207].

Jamwal et al.: Preprint submitted to Elsevier Page 8 of 26

 Survey on Ethereum Pseudonymity

Integrating SMPC in blockchain frameworks epitomizes As depicted in Figure 4, the process involves two parties:
an evolutionary step towards enhancing user privacy [213, a Prover and a Verifier. The Prover sends a confidential
216, 17]. Bitcoin, for instance, has harnessed variations of information, generates a proof of their knowledge using
multi-party computation for the augmentation of transac- func(makeProof), and sends this proof to the Verifier. The
tional privacy, a significant illustration being its deployment Verifier then uses func(checkProof) to verify the proof’s
in the generation of threshold signatures [5]. Central to this validity and concludes the process by obtaining the results.
privacy-centric methodology is the imperative for the ma- This succinctly illustrates the ZKP methodology where the
jority of participants to maintain probity, thereby ensuring Prover can affirm the possession of confidential information
the sanctity and security of the collective computations. without revealing any information itself, thus preserving
Nonetheless, it is pertinent to acknowledge the computa- privacy.
tional intricacies introduced by SMPC, especially in terms
of network latency - attributed to the inherent inter-node data
exchanges requisite for MPC computation [17].
In the decentralized computational landscape, the Enigma
platform, unveiled in 2015, is a paragon of SMPC implemen-
tation [220]. Enigma harnesses a verifiable secret-sharing
scheme, fortifying its computational model’s integrity and
privacy. Further, instead of engendering a nascent blockchain,
Enigma employs an auxiliary blockchain as an immutable
event ledger, concurrently facilitating peer-to-peer network Figure 4: Privacy Preservation techniques based on Zero
governance, which addresses identity management and ac- Knowledge Proof where a formal proof performed by a Verifier
cess control nuances [220]. helps verify program’s execution from a Prover.

3.5. Zero-Knowledge Proof

Zero-Knowledge Proofs (ZKPs) is an advanced cryp- Within the broader category of ZKPs, variations have
tographic technique designed to enhance privacy in the emerged to cater to specific needs. One such adaptation is
blockchain [146]. A relevant use case of ZKP is in iden- the non-interactive variant of ZKP (NIZK) [67, 152]. In
tity verification systems, allowing users to prove specific this NIZK variant, computational zero-knowledge is realized
attributes, like age or citizenship, without revealing other without needing the certifier and verifier to engage directly,
personal details [175]. The origins of this privacy-preserving as long as they both have access to a shared standard ref-
technique trace back to the 1980s when it was proposed as erence string. This finds applications in blockchain, where
a zero-knowledge proof [70, 83]. At its core, the concept all account balances are encrypted on the chain [146, 175].
provides formal proof that verifies a program’s execution When an individual wishes to transfer funds, they can em-
using an input secretly known by the user [82]. These proofs ploy ZKPs to substantiate their ample balance while obscur-
allow one entity to convincingly demonstrate to another ing the actual balance [146].
that it possesses a particular piece of information without Zcash, a ZKP derivative, leverages a type of zero-
divulging any specifics about the data. The execution results knowledge proof called zk-SNARK (Zero-Knowledge Suc-
in a publicly available output, ensuring no sensitive data or cinct Non-Interactive Argument of Knowledge), which con-
attributes are exposed. structs a pool of chaff-coins via the Quadratic Arithmetic
For ZKP to function effectively, it must satisfy three Program [149, 85]. The zk-SNARK proof variation was
critical criteria as follows [27]: first introduced by Bitansky [26] and subsequently formed
the foundation of the Zcash protocol [93]. zk-SNARKs and
1. Completeness: Given a true statement, the prover should its newer version zk-STARKs (Zero-Knowledge Scalable
invariably be able to present a successful proof. Transparent Arguments of Knowledge) offer cryptographic
2. Soundness: If the statement under scrutiny is false, a means to prove knowledge without revealing the information
dishonest prover should find it near-impossible to deceive itself [131, 144]. zk-STARKs is an evolution that provides
the verifier into believing its authenticity, barring a mi- similar benefits to zk-SNARKs without needing a trusted
nuscule probability. setup. They are also post-quantum secure, making them
future-proof against quantum computing threats [144]. The
3. Zero-Knowledge: There should be an algorithm, limited Zcash mechanism employs zk-SNARKs and zk-STARKs to
by polynomial time, that can autonomously create tran- validate transactions, ensuring user privacy remains uncom-
scripts of the protocol which cannot be differentiated promised [149].
from a genuine proof shared between a prover and a
verifier. This is pivotal because if an external entity, 3.6. Differential Privacy
unaware of the statement’s veracity, can fashion a le- Differential privacy, a paradigm conceived by Cynthia
gitimate protocol transcript, neither a verifier nor any Dwork [57], has been acknowledged for pioneering in safe-
interceptor can glean additional information from the real guarding individuals’ privacy in computational processes
interaction. [59]. At the heart of this framework is a mathematically

Jamwal et al.: Preprint submitted to Elsevier Page 9 of 26

 Survey on Ethereum Pseudonymity

rigorous guarantee: an adversary’s knowledge about an in- user information, especially during federated learning pro-
dividual remains invariant, whether or not the individual’s cesses or while chronicling crowd-sourced endeavours [90,
data is part of the computational dataset. Given two adjacent 73, 97, 96].
databases, 𝐷1 and 𝐷2, that differ by precisely one record, A notable stride in this direction was the introduction of a
the probabilities of obtaining any specific output from these blockchain-centric data-sharing framework [73]. This novel
databases should be nearly identical [57, 60]. Consequently, approach empowers data proprietors with the capability to
an adversary remains indecisive about the database of ori- oversee anonymization procedures, thereby thwarting poten-
gin for a particular output, which implies that, even when tial data mining-centric threats targeted at blocking informa-
equipped with supplementary data, the adversary needs to tion. Using local differential privacy (LDP) in conjunction
gain additional insights about the individual in question. with blockchain has also been an approach - LDP is a
The concept is exemplified in Figure 5, which outlines privacy-preserving technique that ensures individual privacy
the fundamental operation of differential privacy. The two while allowing statistical data analysis [59]. By combining
databases, 𝐷1 and 𝐷2, differ by only one record. When blockchain and LDP, a secure genomic data management
each is subjected to an equivalent analysis or computation, system was built that addresses privacy concerns associated
the resulting answers, 𝐴 for 𝐷1 and 𝐵 for 𝐷2, should be with sharing gene data using LDP.[145].
nearly indistinguishable. This ensures that the participation
of an individual’s record in either database does not provide 3.7. Homomorphic Encryption
significant information that could lead to their identification. Homomorphic Encryption (HE), a significant advance-
ment in cryptographic techniques, facilitates operations di-
rectly on encrypted data, obviating the need for decryption
prior to computation [79, 53]. For example, in blockchain-
based healthcare services, HE enables data analytics on pa-
tient records without exposing individual data [189]. There
are primarily two forms of HE: Fully HE (FHE) and Partially
HE (PHE). FHE allows for addition and multiplication oper-
ations on encrypted data, whereas PHE restricts operations
to addition or multiplication [133]. This capability paves the
way for more secure data processing and holds particular
promise for blockchain applications.
These FHE and PHE encryption schemes have been pro-
gressively integrated into blockchain frameworks, fortifying
Figure 5: Differential Privacy safeguards blockchain users’ data data privacy even during computational phases. Such imple-
published onchain with nearly identical outputs when being mentations become indispensable for sectors handling crit-
processed. ical and sensitive data. Within the blockchain domain, exe-
cuting private smart contracts or undertaking computations
The efficiency of differential privacy is parameterized by on confidential data without compromising data integrity
the privacy constant, 𝜀, and the cumulative queries executed is invaluable [113]. Also, in blockchain-based healthcare
over a duration [57, 60, 58]. The smaller the value of 𝜀, services, where patient records demand utmost confidential-
the stronger the privacy guarantee; however, it often comes ity, HE allows data analytics and research without exposing
at a utility cost. Notable applications of differential privacy individual patient data [189, 167].
include its integration in systems like the U.S. Census Bu- Implementing HE in blockchains poses several chal-
reau’s OnTheMap [1], Google’s RAPPOR, Apple, and Mi- lenges due to blockchain systems’ specific needs and proper-
crosoft [105], showcasing its practical usage so far. Moving ties. The computational demands of operating on encrypted
from a theoretical approach to real-world implementation data, mainly using FHE, can hinder the blockchain’s per-
presents challenges, including requiring a skilled person formance and scalability. Designing effective HE schemes
and an appropriate computing development environment and should balance robust security and practical computation
determining the most favourable 𝜀 parameters [78]. with processes to ensure accurate decryption. Integrating en-
In recent years, the blockchain landscape has identified cryption into established blockchain architectures demands
the potential of incorporating differential privacy mecha- attention to compatibility and maintaining the decentraliza-
nisms [125, 90, 73]. For instance, research has been con- tion structure. Beyond technical aspects, there are regulatory
ducted on using differential privacy and blockchain together and legal hurdles since privacy norms differ globally, em-
with Bitcoin [114]. These studies [114] aim to find ways of phasising adherence to global regulations. The complex na-
preserving privacy in the overall structure of the blockchain ture of HE requires enhanced educational awareness and ex-
and to evaluate the effectiveness of differential privacy in im- pertise among blockchain developers and researchers [167].
proving anonymity. Such integrations are paramount when
adversaries exploit blockchain records to deduce sensitive

Jamwal et al.: Preprint submitted to Elsevier Page 10 of 26

 Survey on Ethereum Pseudonymity

4. Deanonymization Techniques on Ethereum cryptocurrencies (i.e., tokens) are linked to specific outputs,
and users own private keys to claim ownership of these
With its open ledger, Ethereum offers unparalleled trans-
outputs. The UTXO model is expected to provide anonymity,
parency, where every transaction is auditable. However, this as the unused transaction information is confidential; it also
transparency does not equate to a complete loss of pri- promotes privacy due to its decentralized transaction linkage
vacy. Users can interact with the Ethereum network through and is amenable to parallelism in transaction validation.
pseudonymous addresses, ensuring transactions are visible
In comparison, the account-based model implemented
while the real-world identity of the actors remains obscured.
by Ethereum is similar to a conventional bank account
Despite employing a number of privacy-preserving tech-
system where each account has an explicit balance. Trans-
niques on Ethereum-based platforms, there is still a possi- actions in this model adjust account balances, leading to
bility that sensitive information about users, including their a direct association between transactions and account ad-
real-world identities, can be inferred and revealed. Ethereum dresses. The design facilitates intricate smart contract func-
has been at the center of several deanonymization attacks,
tionalities and provides a structure that resonates with tra-
ranging from simple address clustering to sophisticated traf-
ditional financial frameworks. In the account-based model,
fic analysis techniques. Ethereum uses accounts as explicit entities. There are two
types of accounts in Ethereum: user accounts, also known
4.1. Account-based vs UTXO Transaction Models
as Externally Owned Accounts (EOA), and smart contracts.
Ethereum employs an account-based transaction model,
Transactions in Ethereum are changes in the account bal-
distinguishing itself from the UTXO model of Bitcoin. As
ance. The account-based model allows for complex state
illustrated in Figure 6, instead of dealing with unspent trans-
transitions through smart contracts written in a Turing-
action outputs, Ethereum maintains a global state of ac-
complete scripting language.
counts where each account possesses an associated balance,
Ethereum’s account-based model and UTXO have dif-
and in the case of smart contract accounts, related code
ferent approaches to handling transactions and account bal-
and internal storage [32]. Transactions in this paradigm
ances. The account-based model in Ethereum allows for
directly alter this global state by transferring Ether between
complex state transitions through smart contracts. In con-
accounts or invoking operations in smart contract accounts,
trast, the UTXO model provides anonymity and can run in
as depicted in the right side of Figure 6. This framework
parallel. Both models have advantages and disadvantages,
facilitates intricate interactions with smart contracts and the
and the choice between them depends on the specific require-
subsequent development of DApps.
ments of the blockchain platform [121].

4.2. Differences in the Deanonymization

Approaches in Ethereum and Bitcoin
Focusing on the architectural difference highlighted in
Figure 6, the deanonymization techniques differ between
Bitcoin and Ethereum due to their foundational transaction
models. As a pure P2P cryptocurrency, Bitcoin’s primary
intent was to offer an electronic cash system free from
centralized intermediaries. Its pseudonymous nature, under-
pinned by unique alphanumeric addresses, ensured that the
entities behind them were not immediately discernible while
transactions were public [4, 46]. Ethereum, proposed in late
Figure 6: Comparison of UTXO vs Account-based Transaction 2013 by Vitalik Buterin [32], is not just a cryptocurrency
Models for cryptocurrencies and/or tokens in Blockchain but a platform for decentralized applications. Unlike Bit-
systems. coin, which was designed primarily for P2P transactions,
Ethereum introduces a more versatile scripting language that
allows for the creation and execution of smart contracts on
The UTXO and the account-based models are founda-
its platform. These contracts are self-executing, with the
tional paradigms that track asset ownership on blockchains.
agreement directly written into lines of code, fostering a
The UTXO model, adopted by Bitcoin and shown on the left
broad spectrum of decentralized applications. This leads to
side of Figure 6, operates on a principle where transactions
more complexity in preserving privacy in Ethereum-based
consume and produce outputs, with the blockchain’s state
systems [18, 31].
being defined by these unspent outputs [134]. For instance,
While both Bitcoin and Ethereum provide pseudony-
in Figure 6, the UTXO model illustrates the changes in the
mous interactions on their respective blockchains, the meth-
list of UTXOs in a blockchain network after each state (e.g.,
ods to deanonymize or link these pseudonyms to real-world
a new block is generated) in which UTXOs are combined
identities vary due to their internal architecture and pur-
and produced (e.g., UTXO2 and UTXO3 are combined in
poses. Bitcoin’s UTXO model and its focus on value trans-
a transaction and UTXO8 is produced accordingly as the
fer have led to specific deanonymization techniques, such
output of the transaction). Technically, in the UTXO model,

Jamwal et al.: Preprint submitted to Elsevier Page 11 of 26

 Survey on Ethereum Pseudonymity

Table 3
Comparison of Deanonymization Techniques: Ethereum vs. Bitcoin.
Technique Bitcoin Ethereum
Transaction Graph Analysis TGA relies on the simpler UTXO model. It TGA is more complex due to the layered
[35, 127] primarily uses the common input ownership interactions involving smart contracts and
heuristic, suggesting that all transaction in- internal transactions. This complexity allows
puts likely belong to one entity. The model for a deeper, though more challenging, anal-
is straightforward and capable of capturing ysis of transaction flows and user behaviours.
complex transaction patterns unique to Bit-
coin.
Address Reuse & Cluster Bitcoin’s vulnerability to address reuse makes Ethereum’s norm of address reuse is similar
Analysis [107, 89] it more amenable to cluster analysis, as to Bitcoin, but the presence of smart con-
addresses can be grouped to identify users tracts acting as mixers introduces additional
potentially. The transparency of its transac- complexity. It requires more sophisticated
tions simplifies this process. methods to differentiate between user and
contract addresses.
Off-Chain Data Cross- In Bitcoin, linking transactions with off- Ethereum’s application of off-chain data
Referencing [81, 107] chain data like KYC(Know Your Customer) cross-referencing is similar to Bitcoin’s. How-
information and social media activity can ever, the involvement of smart contracts and
effectively deanonymize users by associating diverse transaction types adds layers of com-
on-chain activities with real-world identities. plexity in drawing direct associations between
on-chain activities and real-world identities.
Timing Analysis [136, 18] Bitcoin’s timing analysis is effective for pro- Ethereum’s transaction timing offers insights
filing users, as transaction patterns, like fre- into user habits and locations, but the com-
quency and regularity, can be more easily plex nature of transactions, including inter-
correlated with specific behaviours due to the actions with smart contracts, adds additional
regularity and simplicity of Bitcoin transac- layers to the analysis.
tions.
Network Traffic Analysis [42, In Bitcoin, network traffic analysis helps iden- In Ethereum, it is used to study the network’s
24] tify transaction origins and pinpoint node lo- propagation patterns and node distribution.
cations, thus potentially revealing user identi- The diverse node distribution in Ethereum
ties. The analysis benefits from Bitcoin’s less complicates the analysis but helps identify
complex transaction model. security threats and inefficiencies.
Machine Learning [18, 182, Machine learning techniques such as entity Machine learning can analyse transac-
41, 159, 62] identification and graph representation learn- tion data and identify patterns that can
ing are utilized for deanonymization, which deanonymize users by pairing Ethereum ac-
analyzes transaction networks and money counts of the same user and identifying
flows, focusing on identifying and categoriz- sensitive activities linked to public identities.
ing entities and detecting illegal activities.

as the “common input ownership” heuristic [4, 46]. In [173, 187, 102] as shown in 7. A common practice among
contrast, Ethereum’s account-based model and the com- cryptocurrency users is to use a fresh address for every
plexity introduced by smart contracts require a different transaction to protect their privacy. However, when users
approach to deanonymization, focusing more on contract try to consolidate funds from multiple addresses or spend
interactions, internal transactions, and gas usage patterns more than they have in a single address, they unintention-
[18, 31]. Thus, while both blockchains share the broad goal ally link these addresses, allowing for clustering [39, 187].
of pseudonymity, the nuances in their designs have spawned This inadvertent linkage creates identifiable patterns, which,
diverse challenges and methods in forensic analysis and when analysed, can help deanonymize users or at least link
deanonymization [19]. multiple addresses to a single entity [157, 35, 187].
Another aspect to consider is the “change address”. In
4.3. Address Clustering Technique scenarios where the output value of a transaction is less than
Address clustering is one of the most salient techniques its input, a change address is often returned the remaining
for Blockchain, including Ethereum, deanonymization [139, balance. In comparison, Ethereum’s account-based model
63, 39]. The key idea driving this is that addresses used does not function precisely like Bitcoin’s UTXO model, the
as inputs in a single transaction are likely controlled by concepts of fund consolidation and distribution across ad-
the same entity, thereby allowing for the grouping of ad- dresses still apply and can be exploited for clustering. While
dresses into clusters representing individual users or entities address clustering can provide valuable insights, it is crucial

Jamwal et al.: Preprint submitted to Elsevier Page 12 of 26

 Survey on Ethereum Pseudonymity

to approach its results with caution. False positives, where millions of Ethereum transactions to identify blacklisted ad-
unrelated addresses are incorrectly linked, can occur. This dresses and potential security threats within seconds [110].
mandates using supplemental information or techniques to
validate the associations derived from clustering.

Figure 8: Transaction Graph Analysis technique used PageR-

ank metrics over the volumes of transactions to characterise
Figure 7: Address clustering technique groups some addresses different types of nodes in Bitcoin network
that might belong to a user (i.e., e1 or e2)

In the context of Ethereum, tools like Tornado Cash 4.5. Transaction Fingerprint and Timing Analysis
have emerged to enhance transactional privacy [147, 109]. Transaction fingerprint is a specialised technique used
Tornado Cash, a non-custodial coin mixer on Ethereum, predominantly in blockchain analytics, focusing on some
is designed to protect the privacy of addresses [147, 109]. unique attributes of transactions, including the temporal
However, certain transaction behaviours within its mixing attributes such as Hour of Day, Time of Day, Time of Hour,
mechanism can lead to potential privacy risks. Specifically, and Random time-interval so-called Timing Analysis [4,
malicious attackers can link multiple addresses of the same 18]. Generally, by analyzing the timestamps of transactions
users based on transaction data. This highlights the chal- within a blockchain, timing analysis endeavours to identify
lenges and nuances of maintaining privacy even when us- patterns, correlate activities, and potentially deduce the iden-
ing additional privacy-preservation techniques designed to tities or behaviours of participants [136]. The foundational
enhance it [178]. theory of the timing analysis in blockchain revolves around
the observation that while addresses may be pseudony-
4.4. Transaction Graph Analysis mous, their activities, marked by timestamps, are transparent
Transaction Graph Analysis (TGA) leverages graph the- [4, 18]. This transparency means that by merely analyzing
ory to map out the intricate web of blockchain transac- the frequency, regularity, or timing of transactions from a
tions. Nodes correspond to individual addresses, and di- specific address, it might be possible to glean information
rected edges trace the movement of currency, revealing about the user or entity behind it. For instance, patterns
transaction volumes and patterns [107]. For instance, TGA that emerge, such as transactions consistently occurring at
is used to depict a typical daily transaction graph where specific times of the day, can provide hints about the user’s
nodes, indicated by PageRank metrics, represent critical timezone or habits [18].
actors within the blockchain as illustrated in Figure 8. These Other transaction fingerprints can also be Coin-Flow
nodes significantly influence the flow of cryptocurrency due and Input/Output balance [4] so that when combined with
to their high transaction volumes. This method is integral the temporal attributes, it might increase the probability to
to detecting typical and anomalous behaviour within the deanonymize individual users. For instance, Androulaki et
network, which is essential for security analysis and market al. have based on the transaction fingerprints of Bitcoin
research [11, 127]. transactions in a university campus to successfully identify
TGA has become a pivotal tool in privacy research, 40% user profiles, even when the users generate new ad-
uncovering the blockchain’s pseudonymity limits and iden- dresses for every transaction. This technique can be imple-
tifying traceable patterns that could lead back to real-world mented for Ethereum and Bitcoin transactions, regardless
entities [4, 35, 72, 158]. It faces challenges from evolving of their transaction models (i.e., account-based and UTXO)
blockchain features, like smart contracts, and obfuscation [136, 18].
methods such as tumblers [74, 42, 103]. In addressing se- A notable application of the transaction fingerprint tech-
curity concerns, graph analysis has revealed the identities nique is the study of user behaviours concerning exchange
behind Ethereum attacks by dissecting transaction data asso- platforms [181]. For example, when a user interacts with a
ciated with various security breaches [35, 42, 119, 210, 100]. decentralized exchange, there is often a discernible delay
Advances in TGA have enabled the rapid analysis of exten- between when funds are sent to an exchange address and
sive data, such as in the work of researchers who analyzed when they are withdrawn or traded. By closely examining

Jamwal et al.: Preprint submitted to Elsevier Page 13 of 26

 Survey on Ethereum Pseudonymity

these time deltas and cross-referencing them with off-chain blockchain analysis. By analyzing transaction-related textual
events like exchange downtimes or high trading volumes, data and public forum discussions, NLP techniques help
researchers can get insights into user strategies, preferences, correlate on-chain activities with real-world identities [159],
or even identities in some cases [14, 181]. which could add another dimension to deanonymization
Despite its usefulness, this analysis comes with certain methodologies. Predictive analytics using historical data
limitations. The increasing adoption of privacy-enhancing could forecast future transaction patterns and behaviours
technologies and best practices, like CoinJoin or mixers, in blockchain, identifying unusual or user-specific patterns.
can obfuscate transaction patterns, making pure timing- These insights are crucial to uncovering the hidden aspects
based analysis less effective. Furthermore, as blockchains of pseudonymity in blockchain, offering a deeper under-
grow and transaction volumes swell, the sheer number of standing of user behavior that necessitates further study
transactions can make it challenging to definitively correlate [41, 18, 159].
events solely based on timings [155]. This section has elucidated various deanonymization
techniques, each with a unique facet of potential vulnerabili-
4.6. Network-Level Analysis ties within the Ethereum network. From address clustering’s
Blockchain operations depend on a peer-to-peer net- propensity to link multiple addresses to transaction graph
work, making understanding this layer crucial for improv- analysis’s power to reveal intricate transactional relation-
ing system efficiency and examining vulnerabilities. Since ships to the subtle cues offered by timing analysis and the
blockchain interactions leave traces at the network layer, holistic insights of network-level examinations, each method
their study can reveal insights into on-chain pseudonymity offers a glimpse into the hidden intricacies of user interac-
[92]. The network-level analysis investigates interactions, tions. However, these techniques are not without challenges,
data flows, and traffic in blockchain networks, seeking in- demanding continuous refinement in light of the evolving
formation about nodes, transactions, and user behaviour blockchain ecosystem.
[25, 23]. Central to this analysis is traffic pattern examination
[24]. When a transaction is broadcast, it initially reaches
a few nodes before spreading network-wide. Identifying a 5. Countermeasures for Deanonymization on
transaction’s origin or early receivers helps pinpoint the Ethereum platforms: the Road Ahead
initiator. Analyzing block propagation can hint at miners’ Although revolutionary for its decentralized and im-
locations or identities. mutable characteristics, blockchain technology confronts
This approach helps identify Sybil attacks, where at- significant privacy preservation challenges. The public na-
tackers create many nodes to disrupt traffic [193]. Recent ture of permissionless blockchain ensures transparency but
research has shed light on a significant privacy concern: the may also inadvertently expose some sensitive information.
exposure of Bitcoin users to vulnerabilities through Tor’s Ethereum is not an exception, even though it is renowned
hidden services [2]. Tracing a Bitcoin transaction back to for its pseudonymous nature, as the platform maintains
its origin is feasible by scrutinising publicly available data. public transactions without directly revealing real-world
Linking a previous transaction to a specific Bitcoin address identities—the trade-off between its pseudonymous nature
can reveal the user’s identity. This revelation highlights and the risk of exposing sensitive transactional information.
a critical intersection between blockchain technology and Privacy preservation in permissionless blockchain systems
user privacy, underlining the need for enhanced security like Ethereum is a pivotal yet intricate endeavour, and
measures in blockchain networks to protect user anonymity achieving the ideal balance between transparency, security,
[2]. and privacy still presents a multifaceted set of challenges.
In this section, we delve into the complex landscape of pri-
4.7. Machine Learning-based Analysis
vacy preservation in permissionless blockchain systems like
Machine learning has revolutionized blockchain analysis
Ethereum, exploring the intricate challenges and potential
by enabling the identification of complex patterns within
solutions.
transaction data. These algorithms analyze large datasets to
uncover transaction frequencies, amounts, timing, and in- 5.1. Openness versus Privacy: a Long-standing
teractions between addresses, aiding in profiling blockchain
Challenge
users and potentially linking pseudonymous addresses to
Ethereum, often complimented for its transparency and
real-world identities [18]. This capability extends to iden-
decentralized nature, inherently poses significant challenges
tifying control over multiple addresses by single entities,
when ensuring user privacy. This subsection discusses the
which is crucial for deanonymization efforts. Complement-
inherent conflict between Ethereum’s design priorities and
ing this, graph databases provide a powerful means to vi-
privacy concerns. The system design and techniques imple-
sualize and analyze the interconnectedness of transactions
mented in the Ethereum blockchain prioritise transparency
and addresses, unveiling fund movement and interaction
over anonymity. This intrinsic design ensures trustworthi-
patterns that are key to understanding blockchain dynamics
ness among decentralized nodes, as every transaction is ver-
[75, 18, 41].
ifiable by any participant. However, the flip side to openness
In parallel, Natural Language Processing (NLP) and
is that all transaction details become publicly available. Even
predictive analytics contribute significantly to the depth of

Jamwal et al.: Preprint submitted to Elsevier Page 14 of 26

 Survey on Ethereum Pseudonymity

though Ethereum offers pseudonymity to an extent through Table 4

unique cryptographic addresses, they should go further than Understanding the multifaceted complexities of Smart Con-
that to provide better privacy. Balancing the scale between tracts development in Ethereum.
openness and privacy preservation becomes a complex en- Complexity
deavour deeply rooted in the foundational principles of tech- Description
Types
nology. However, this transparency comes at the cost of Computational resources consumption.
Computational
privacy, as detailed below. It is associated with gas costs, requiring
Complexity [41]
optimisation for cost efficiency.
• Immutable Public Ledgers: At the core of permissionless Rigorous validation for Smart Con-
Logical
blockchains like Ethereum is a publicly available im- tracts as any mistakes can lead to
Complexity
mutable record of all transactions. While this design en- [22, 10]
unintended consequences, including fi-
sures transparency and trustworthiness, it exposes trans- nancial losses.
action details to the public [150]. Finding ways to conceal Inter-contract Dependencies among Smart Contracts
specific transaction data without compromising ledger Complexity emerging complexity as a Smart con-
integrity is challenging. [101] tract can invoke others.
Usability
Requirement for user-friendly inter-
• Transparency in Smart Contracts: Ethereum’s smart con- Complexity
faces with abstraction layers.
tracts, in enabling programmable transactions, are visible [45]
on the blockchain. The contract’s transparency can inad-
vertently expose private logical methods or data. The data
associating Ethereum addresses with identifiable data on
in these contracts can also be publicly accessed [106].
external platforms can expose user identities. Anonymity
For example, suppose a smart contract is designed to
is the form of privacy that we are targeting. Unfortunately,
store user data or credentials. This data is visible on the
it comes with some far-reaching implications.
blockchain unless specific privacy-enhancing techniques
are employed, like zk-SNARK [149, 85]. • Scalability and Privacy Trade-offs: Privacy-preserving
The transparency of the Ethereum blockchain allows any- cryptographic techniques, like zk-SNARKs, which are
one to view all transactions and smart contract interac- utilized in specific Ethereum-based applications, pose
tions. This lack of privacy can be a barrier for individuals substantial computational and storage usage [131]. Incor-
and organizations that require confidentiality for their porating these methods can lead to increased block veri-
transactions [69]. Several privacy-enhancing technolo- fication times and higher gas fees. Ethereum developers
gies, such as mixers [126, 147, 166], stealth addresses would have to weigh the benefits of enhanced privacy
[47], and confidential transactions[31], have been pro- against the potential drawbacks regarding system perfor-
posed and deployed to enhance blockchain-level privacy. mance and transaction efficiency.
These technologies aim to obfuscate transaction details
• Interoperability Concerns: The drive towards blockchain
and protect participants’ privacy on the blockchain.
interoperability—different chains communicating and trans-
5.2. Blockchain-related Technical Challenges actions pose privacy concerns. A transaction might main-
Some technical challenges hinder the efficiency and ef- tain privacy within one chain but risk exposure during
fectiveness of privacy preservation techniques in blockchain- cross-chain operations.
based systems due to the nature of permissionless blockchain
platforms. These challenges are always must-considered fac-
5.3. Complexity of Smart Contracts
Smart contracts, as implemented in blockchain platforms
tors for any privacy preservation techniques, which include
such as Ethereum, have introduced a paradigm shift in
(i) the balance between anonymity and pseudonymity, (ii)
executing contractual agreements [32]. Their inherent nature
the trade-off between the high cost of computation and
enables automation, trust, and decentralization. However, as
storage, the scalability and the user privacy, and (iii) the
shown in Table 4, with these advantages come complexities
interoperability between on-chain and off-chain computation
both in their design and operations, which result in complica-
and storage.
tions in guaranteeing security and privacy effectively [218].
• Pseudonymity vs Anonymity: Ethereum operates on a pub-
• Computational Complexity: Every operation in a smart
lic ledger where transactions are tied to pseudonymous
contract, from arithmetic operations to storage, consumes
addresses. These addresses, commonly denoted as a 42-
a certain amount of computational resources. Especially
character hexadecimal string (e.g., 0𝑥...), serve as an iden-
in platforms like Ethereum, these operations are associ-
tifier on the Ethereum network. Unlike real-world identi-
ated with gas costs [41]. Optimising the computational
ties, these addresses provide a degree of pseudonymity but
tasks within a contract is paramount for cost efficiency
not anonymity (i.e., absolute privacy) [77]. As presented
and ensuring the contract does not run out of gas, which
in Section 4, vulnerabilities like pattern and behavioural
would result in a failed transaction.
analysis over the Ethereum transaction graph, repeated
usage of Ethereum addresses in various interactions, and

Jamwal et al.: Preprint submitted to Elsevier Page 15 of 26

 Survey on Ethereum Pseudonymity

Figure 9: Forthcoming solutions for enhancing privacy on the Ethereum blockchain, encompassing mixers, P2P security, Layer-2
scaling, and educational initiatives.

• Logical Complexity: Ensuring a contract behaves as in- 5.4.1. Use of Mixers

tended requires rigorous validation, given the immutable Ethereum mixers either function as third-party services
nature of blockchains [22]. Once deployed, a contract’s (i.e., centralised mixer, generally integrated with an ex-
logic cannot be altered. Mistakes can lead to unintended change) or a decentralized mixer (i.e., operated at the Ap-
consequences, including financial losses [10]. plication layer in the Ethereum layered stack as depicted in
Figure 9) that help users obfuscate their transactions. By
• Inter-contract Complexity: Smart contracts can invoke sending their Ether to the mixer, users receive another set
other contracts, creating a web of dependencies [101]. of Ether from a different address that cannot be directly
This can lead to emergent complexities that are harder to linked to their original one. Decentralized mixers seem to
predict at the time. be a prospective solution, and several cryptocurrencies have
• Usability Complexity: For non-technical users, smart con- developed this approach to their protocols, such as Tornado
tracts’ intricacies, deployment, and management can be Cash [147] and Zether [31].
overwhelming [45]. User-friendly interfaces and abstrac- Tornado Cash is a decentralized Ethereum mixer that
tion layers are essential to make smart contracts accessible uses zk-SNARKs to ensure that the outgoing transaction
to a broader audience. cannot be linked to the corresponding deposit, thereby break-
ing the on-chain link between the source and destination
5.4. Forthcoming Solutions addresses [147]. Zether is a fully decentralized, confiden-
Potential innovative countermeasures for deanonymiza- tial payment mechanism compatible with Ethereum’s con-
tion have breathlessly been developed in the Ethereum sensus model. It uses zero-knowledge proofs, specifically
ecosystem. These emerging solutions range from advanc- Bulletproofs, to maintain transaction confidentiality [31].
ed cryptographic techniques to off-chain strategies, from Bulletproofs are a non-interactive zero-knowledge proof sys-
network-level security to system scalability while strength- tem that greatly reduces computational overhead and proof
ening user privacy and security. We will delve into using sizes. In comparison against systems like zk-SNARKs [131],
mixers, network-layer security improvements, Blockchain Zether does not require a trusted setup. This presents a
layer-2 scaling techniques, and educational initiatives that considerable advantage as such setups can become critical
collectively contribute to the Ethereum blockchain’s ro- vulnerabilities if they were to be compromised.
bustness and resilience. As illustrated in Figure 9, these
countermeasures span multiple stakeholders, including wal- 5.4.2. P2P Network-layer Solutions
let service, centralised third-party service, and the layered The P2P layer of blockchain technology presents unique
Ethereum platform (e.g., Application, Layer-2, Consensus, challenges and opportunities for user anonymity, particularly
and P2P Network Layers). within the Ethereum ecosystem. A deeper analysis employ-
ing network and data traffic indicators reveals a pressing
need to deploy sophisticated privacy-preserving techniques
rapidly [77]. Integrating the Tor network with blockchain

Jamwal et al.: Preprint submitted to Elsevier Page 16 of 26

 Survey on Ethereum Pseudonymity

technology significantly enhances anonymity and exposes integration facilitates the aggregation of hundreds of trans-
distinct packet characteristics that may compromise privacy actions into one, significantly enhancing scalability and pri-
[92]. vacy on the Ethereum blockchain [117]. One of the notable
Recent advancements in P2P protocols demonstrate Ethereum’s zk-SNARKs integration methods, Ethereum’s
the potential to obfuscate the origin and destination of Istanbul upgrade, has brought notable changes in gas costs,
blockchain transactions, thereby strengthening anonymity allowing for zk-SNARKs operations to be more affordable
[148]. These advancements include onion routing [40, 104] and hence paving the way for more private transactions on
and cryptographic shuffling [111], instrumental in con- Ethereum [131].
cealing user 𝐼𝑃 addresses and transaction details from Another type of Rollups is Optimistic Rollups tech-
network observers. Such techniques are imperative in an nique, which posts the transaction data to the main-chain
era where blockchain forensic tools are increasingly adept but executes them on the Layer-2 network [7]. Unlike zk-
at deanonymizing P2P ledger activities. Rollups, this Optimistic Rollups technique does not use
The advocacy for anonymous communication channels, ZKP-variants; instead, it assumes transactions are valid by
particularly the Tor network, emphasizes the necessity of default and only runs computations in case of a challenge us-
access-related privacy within decentralized architectures. ing its dispute resolution mechanism [7]. To do so, optimistic
The scholarly consensus points towards the need for systems rollups implement a fraud-proving scheme to detect incor-
that enable users to conduct and retrieve transactions without rect transactions by computing a fraud-proof for the chal-
revealing network addresses or transaction specifics. This lenge [7, 163]. Ethereum co-founder’s belief in rollups as a
requirement extends to access transaction details without di- promising solution for Layer-2 functionality in Ethereum 2.0
vulging individual transactional behaviours or preferences, underscores their importance in addressing scalability issues
preserving privacy against traffic analysis attacks [92, 148]. while maintaining data privacy and availability [171, 196].
Integrating privacy-focused P2P solutions with Ethere- A new development in the Ethereum ecosystem called
um’s smart contracts and DApps can address some of the in- Proto-Danksharding is an evolution of sharding technology
herent limitations of public blockchains in protecting trans- to improve network scalability. It is designed to increase the
actional privacy. Such integrations could also facilitate a space available for processing large volumes of data, thereby
more secure and private infrastructure for decentralized fi- enhancing the overall efficiency of the Ethereum network.
nance (DeFi) applications, which are particularly sensitive This approach is expected to reduce gas fees significantly
to the risks of financial privacy breaches. and speed up transaction processing, making Ethereum more
scalable and capable of supporting a more extensive user
5.4.3. Layer-2 Solutions base [68].
Layer-2 scaling solutions address the scalability and
privacy challenges inherent to Ethereum’s Layer-1. These 5.4.4. Off-chain Solutions
solutions operate on top of the Ethereum mainnet (Layer- Off-chain solutions allow transactions to be processed
1). Being Layer-2 solutions, these methods operate on top at a third-party centralised server, outside the Ethereum
of the Ethereum mainnet and can process transactions more main-net, with only the final state or a summary updated
efficiently, offering scalability and enhanced privacy [117]. on the main chain [112]. The off-chain solutions primarily
One of the prominent Layer-2 solutions is Rollups, a cooperate with the blockchain main-net at Layer-2 to record
technique that uses Layer-2 network entities to process com- the final state or updates on-chain using several Layer-2
ponent transactions outside of the main-chain. These entities on-chain/off-chain interaction techniques, including State
then interact with the blockchain through smart contracts to Channels and Plasma. This approach offers significant ad-
post transaction data to the main chain as calldata. Rollups vantages in terms of both scalability and privacy [112].
provide a mechanism to batch multiple transactions into one
transaction, reducing the on-chain data footprint. By aggre- • State Channels: State channels are two-way pathways
gating multiple transactions, Rollups enhance the through- opened between two participants. Once a state channel
put and make it more challenging for adversaries to analyze is opened, participants can transact amongst themselves
individual transactions, offering privacy [180]. without involving the main chain [61]. Only their transac-
There are two primary types of Rollups: zk-Rollups and tions’ final state is recorded on the Ethereum blockchain.
Optimistic Rollups [180]. zk-Rollups technique utilizes zk- This ensures rapid and cost-effective transactions and
SNARKs to bundle multiple transactions into a single proof, provides privacy, as intermediate states are not publicly
which is then submitted to the main-chain. This ensures recorded. State channels benefit applications where mul-
scalability and provides a layer of privacy due to the suc- tiple rapid transactions are expected between participants,
cinct nature of zk-SNARKs [135]. The integration of zk- such as in gaming or micro-payment scenarios [191].
SNARKs, a privacy cryptography protocol, and zk-Rollups,
• Plasma: Plasma is a framework that allows for the cre-
a Layer-2 scaling solution, plays a crucial role in improving
ation of child chains branching off the main Ethereum
the scalability and privacy of the Ethereum ecosystem. This
chain. These child chains can process transactions in-
dependently, reducing the load on the main Ethereum

Jamwal et al.: Preprint submitted to Elsevier Page 17 of 26

 Survey on Ethereum Pseudonymity

chain [151]. Periodically, these child chains submit a sum- 6.1. Regulatory-compliance Challenges
mary of their transactions to the main chain. Since only Integrating blockchain technology, specifically Ethere-
summaries are recorded on the main chain, individual um’s decentralized architecture, with the GDPR presents a
transaction details on the child chains remain obfuscated, complex and challenging landscape [71]. At the core of this
providing a layer of privacy [151]. tension lies the differing operational philosophies: Ethereum
operates on principles of decentralization and immutability,
5.4.5. Education and Privacy Awareness whereas GDPR is built around individual data rights and
As seen in platforms like Ethereum, the rapid advance- control. This dichotomy raises pertinent questions regarding
ment of blockchain technology emphasizes the importance assigning roles like data controllers and processors within
of education and privacy awareness when navigating a bal- a decentralized system, a fundamental aspect of GDPR’s
ance between decentralization, pseudonymity, and regula- framework [176].
tory compliance [154]. Although Ethereum provides a layer
of privacy through its inherent pseudonymity, it does not of- • Legal and Regulatory Complexity: Blockchain technol-
fer complete anonymity. Users must be well-informed about ogy’s cornerstone is its immutability. Once data enters
the nuances of their public addresses, transaction patterns, the blockchain, its alteration or deletion becomes impos-
and the potential for linkage to real-world identities to make sible. This characteristic directly challenges Article 17 of
informed decisions about their on-chain activities. the GDPR, which introduces the “right to be forgotten,”
As deanonymization techniques evolve, raising aware- granting individuals the privilege to request their personal
ness about these methods within the Ethereum community data’s deletion under predefined conditions [64]. Appli-
is crucial. Workshops, webinars, and online resources can be cations developed atop the Ethereum platform, which
vital in disseminating knowledge about emerging threats and inscribe personal data directly onto the blockchain, may
countermeasures. While smart contracts have the potential inherently breach the GDPR’s right to erasure. This has
to revolutionize contractual paradigms, they also introduce led to extensive deliberations among technologists and
unique privacy challenges. Users require guidance on vetting legal scholars [94]. sqlCopy code
smart contracts, understanding their underlying logic, and
• Regulatory and Compliance Pressures: Regulatory bod-
ensuring their interactions do not compromise personal data.
ies worldwide are increasingly interested in cryptocurren-
cies. Though often well-intended, their push for oversight
6. Regulatory-compliance for DApps and transparency can place strains on maintaining user
privacy. For instance, as smart contracts can represent and
In April 2016, an important data protection regula-
even automate real-world agreements, they often intersect
tion, the GDPR (General Data Protection Regulation), was
with traditional legal systems [194]. Determining the ju-
adopted. It became enforceable in May 2018 and was created
risdiction, applicability of laws, and the legal status of a
by the European Parliament, the Council of the European
smart contract execution **presents unique** challenges.
Union, and the European Commission [64]. While GDPR
primarily targets data protection within the EU and EEA, its
6.2. Regulatory Implications
global implications affect any organization dealing with EU
The decentralized nature of Ethereum, where multiple
residents’ data. GDPR aims to consolidate data protection
nodes validate and record transactions simultaneously, dis-
for all individuals within the European Union (EU) and
rupts the traditional understanding of the three roles (i.e.,
the European Economic Area (EEA), emphasizing giving
Data Subject, Data Controller, and Data Processor) defined
citizens and residents greater control over their data. Addi-
in the GDPR. This dissonance complicates the identification
tionally, it aims to streamline the regulatory framework for
of responsible parties (e.g., lawful bases for data processing,
international businesses [64].
the obligations of the Data Controller, and the rights of
The utilisation of blockchain architecture has increased
Data Subjects). It also raises concerns about accountability
since Nakamoto publicised the foundational article in 2009,
in cases of regulatory breaches. Below are notable aspects
particularly after the inauguration of Ethereum in 2015
considered for harmonizing permissionless blockchains like
[134]. Since then, many solutions have relied on this de-
Ethereum with the GDPR [176].
centralized database to process personal data. However,
the permanent nature of recorded data in these solutions • Right to be Forgotten: The GDPR, through Article 17, in-
raises concerns about how to ensure GDPR compliance. troduces the “Right to be Forgotten”, allowing individuals
The challenge intensifies when considering the global na- the power to have their personal data erased under certain
ture of blockchain technology and the varying data protec- conditions [64]. This clashes directly with the immutabil-
tion laws across different jurisdictions. This paper examines ity feature of blockchain. It was mentioned that the inher-
the regulatory-compliance challenges and their implications ent unsuitability of decentralized systems like blockchain
arising from the intersection of the GDPR and Ethereum, for storing personal data [80]. Instead of hosting personal
focusing mainly on areas of compatibility and tension. data, blockchain’s strengths should be towards achieving
consensus and verification, with personal data storage
allocated to traditional systems. This aligns with Bitcoin’s

Jamwal et al.: Preprint submitted to Elsevier Page 18 of 26

 Survey on Ethereum Pseudonymity

initial design, which operates without storing directly method, however, may conflict with the decentralized
identifiable data. By using Ethereum addresses instead of nature of blockchain [170].
personal identification, the public ledgers sidestep many In healthcare, where GDPR compliance is critical, block-
GDPR issues. chain applications must carefully navigate data protection
A solution to reconcile the GDPR with blockchain is and interoperability requirements. Ensuring patient data
through encryption [80]. It was pointed out that encrypted transferred across borders via blockchain technologies
personal data can fulfil GDPR’s erasure requirements if complies with GDPR is essential [91].
the decryption key is held off-chain or destroyed. The
idea of a referenced approach was presented [183]. Var- • Privacy by Design (PbD): The principle of “Privacy by
ious GDPR compliance proposals for blockchain include Design” (PbD) is a proactive approach for centralised
Chameleon Hash for data modification, secret sharing for systems, suggesting that privacy considerations should be
more comprehensive modification access, cryptographic central to any system or process design from its inception
deletion via Linkable Digital Multi-signature, minimal- [120]. The GDPR has enshrined this concept, making it
impact alteration with Reconstructable Ephemeral Key, a legal requirement for any system handling the personal
local Erasure Database for participant-led data removal, data of EU citizens as the role of a Data Controller [183].
blockchain-based authentication and authorization, and In DApps developed on blockchain platforms such as
Tree Structure for more straightforward data deletion [80]. Ethereum, the inherent decentralization characteristic of
these platforms is mirrored. This architectural choice
• Automated Decisions and Smart Contracts: Smart con- negates centralized control over the operations of DApps,
tracts, conceptualised by Szabo, execute contractual obli- leading to the distribution of operational data across many
gations automatically [177]. However, this concept en- nodes. This distribution paradigm presents significant
counters GDPR’s Article 22, which grants individuals the deviations from traditional data management and privacy
right to challenge decisions based exclusively on auto- protocols [94, 183, 219]. Recent scholarly discussions
mated processing. The significance of integrating ’exit have pointed out that the attributes underpinning the
strategies’ within smart contracts to ensure GDPR com- revolutionary nature of DApps may concurrently result in
pliance has been underscored in the past. These strate- inherent challenges in aligning with GDPR compliance
gies are essential for adapting smart contracts to legal [219]. Establishing a centralized oversight entity within
frameworks like the GDPR, allowing for adjustments in the decentralized sphere of DApps is a potential solution.
response to automated decisions [51]. Consent remains This entity would be tasked with monitoring, enforcing,
a significant element in the GDPR’s framework. Trans- and rectifying potential GDPR breaches [165]. It could
parent and explicit consent in blockchain contexts could also serve as a resource for standardized guidelines, best
mitigate potential conflicts. This approach is particularly practices, and tools, thereby ensuring the embedding of
effective when users can challenge or modify the conse- privacy considerations at the core of DApps’ design under
quences of their consent. the Privacy by Design principle.
• Cross-border transfer: The General Data Protection Reg- While Ethereum’s principles may initially seem antithet-
ulation (GDPR) has specific provisions concerning trans- ical to GDPR’s mandates, exploring avenues that harmo-
ferring personal data across borders, particularly outside nize these conflicting foundations is imperative, particularly
the European Economic Area (EEA). Article 44 of the in incorporating new technologies into existing regulatory
GDPR [64] mandates that any transfer of personal data frameworks. This reconciliation is vital for ensuring the
which are undergoing processing or are intended for pro- coexistence and advancement of groundbreaking blockchain
cessing after transfer to a third country or an international capabilities and protecting user data, including ensuring ac-
organization shall take place only if the conditions laid countability of data processing as well as offering individual
down in the regulation are complied with by the data con- data rights, as one of the ultimate goals for establishing the
troller and processor, ensuring that the level of protection GDPR.
guaranteed by the GDPR is not undermined [64].
In the blockchain and DApps context, the distributed 7. Conclusion
nature of blockchain technologies presents challenges in
ensuring GDPR compliance. Anonymizing personal data This insightful survey paper has comprehensively ex-
in blockchains through public key cryptography may not plored the intricate dynamics of privacy and deanonymiza-
exempt it from GDPR, as pseudonymous data still qualify tion in the Ethereum blockchain ecosystem. The decen-
as personal data under the regulation [150]. Adapting tralized and open-source nature of Ethereum, while instru-
blockchain to comply with GDPR’s data transfer require- mental in driving innovation and transparency, concurrently
ments may require new legal frameworks or significant poses significant privacy challenges. The inherent trans-
modifications to existing ones [88]. One approach is ’data parency of Ethereum’s public ledger, essential for trust and
localization’, which involves storing and processing data auditability, paradoxically creates avenues for potential pri-
within the data origin’s geographical boundaries. This vacy invasions. Deanonymization techniques, such as ad-
dress clustering, transaction graph analysis, and transaction

Jamwal et al.: Preprint submitted to Elsevier Page 19 of 26

 Survey on Ethereum Pseudonymity

fingerprint analysis, leverage the openly available transac- 2013, Okinawa, Japan, April 1-5, 2013, Revised Selected Papers 17,
tional data to potentially expose user identities and be- Springer. pp. 34–51.
haviours. Furthermore, these techniques can be combined [5] Andrychowicz, M., Dziembowski, S., Malinowski, D., Mazurek, Ł.,
2016. Secure multiparty computations on bitcoin. Communications
with off-chain information including network-level, social of the ACM 59, 76–84.
network, and web-crawling data for more effectiveness of [6] Antonopoulos, A.M., Wood, G., 2018. Mastering ethereum: building
the re-identification of users. While beneficial for forensic smart contracts and dapps. O’reilly Media.
analysis and network security, these techniques underscore [7] Armstrong, M., 2021. Ethereum, smart contracts and the optimistic
the vulnerabilities in Ethereum’s pseudonymous framework. roll-up. BSc Dissertation: University of Dublin, Trinity College .
[8] Asplund, A., Hartvigsen, P.F., 2015. Reclaiming data ownership:
The advent of various privacy-preserving methodolo- Differential privacy in a decentralized setting. URL: https://api.
gies, including mixers, zk-SNARKs, and network-layer so- semanticscholar.org/CorpusID:55245471.
lutions, signifies the ongoing efforts to bolster privacy in [9] Aspnes, J., Attiya, H., Censor, K., 2010. Combining shared-coin
Ethereum. However, these solutions have limitations, often algorithms. Journal of Parallel and Distributed Computing 70, 317–
with trade-offs between privacy, scalability, and regulatory 322.
[10] Atzei, N., Bartoletti, M., Cimoli, T., 2017. A survey of attacks on
compliance. Introducing blockchain layer-2 solutions, off- ethereum smart contracts (sok). Principles of Security and Trust
chain transactions, and advanced privacy enhancing tech- 10204, 164–186.
niques like zk-STARKs presents promising avenues for user [11] Bai, Q., Zhang, C., Liu, N., Chen, X., Xu, Y., Wang, X., 2021.
privacy. However, they also necessitate a balance between Evolution of transaction pattern in ethereum: A temporal graph
the inherent decentralization of blockchain and the regula- perspective. IEEE Transactions on Computational Social Systems
9, 851–866.
tory demands, particularly in light of GDPR compliance. [12] Banerjee, S., Das, D., Biswas, M., Biswas, U., 2020. Study and
Furthermore, the complexity of smart contracts adds another survey on blockchain privacy and security issues, in: Cross-industry
layer to the privacy puzzle, necessitating robust design and use of Blockchain Technology and Opportunities for the Future. IGI
operational strategies to safeguard user data. Global, pp. 80–102.
The tension between the immutable nature of blockchain [13] Baria, N., Parmar, D., Panchal, V., 2022. Blockchain user, network
and system-level attacks and mitigation, in: The Auditor’s Guide to
and emerging data protection regulations like the GDPR also Blockchain Technology. CRC Press, pp. 223–243.
presents a formidable challenge, calling for innovative solu- [14] Baum, C., David, B., Frederiksen, T.K., 2021. P2dex: privacy-
tions that harmonize decentralized technology with evolving preserving decentralized cryptocurrency exchange, in: Interna-
legal frameworks. Presumably, Ethereum’s blockchain sig- tional Conference on Applied Cryptography and Network Security,
nificantly advances decentralized applications; however, its Springer. pp. 163–194.
[15] Benabdallah, A., Audras, A., Coudert, L., El Madhoun, N., Badra,
engagement with privacy issues and the need to comply with M., 2022. Analysis of blockchain solutions for e-voting: A system-
regulatory standards is an ongoing and developing area. atic literature review. IEEE Access .
Therefore, the road ahead should concentrate on creating [16] Benet, J., 2014. Ipfs-content addressed, versioned, p2p file system.
advanced methods to protect privacy in ways that meet arXiv preprint arXiv:1407.3561 .
legal requirements, improve the security of smart contracts, [17] Benhamouda, F., Halevi, S., Halevi, T., 2019. Supporting private
data on hyperledger fabric with secure multiparty computation. IBM
and promote better understanding and knowledge about Journal of Research and Development 63, 3–1.
these complex issues. Achieving a balance between trans- [18] Béres, F., Seres, I.A., Benczúr, A.A., Quintyne-Collins, M., 2021.
parency, privacy, and adherence to regulatory frameworks Blockchain is watching you: Profiling and deanonymizing ethereum
in Ethereum goes beyond technology alone; it also involves users, in: 2021 IEEE international conference on decentralized ap-
legal and ethical considerations. This necessitates a collabo- plications and infrastructures (DAPPS), IEEE. pp. 69–78.
[19] Bernabe, J.B., Canovas, J.L., Hernandez-Ramos, J.L., Moreno, R.T.,
rative effort across various disciplines to ensure blockchain Skarmeta, A., 2019. Privacy-preserving solutions for blockchain:
technology’s responsible and sustainable growth. Review and challenges. IEEE Access 7, 164908–164940.
[20] Bertino, E., Kundu, A., Sura, Z., 2019. Data transparency with
blockchain and ai ethics. Journal of Data and Information Quality
References (JDIQ) 11, 1–8.
[1] Abowd, J.M., 2018. The us census bureau adopts differential privacy, [21] Bethencourt, J., Sahai, A., Waters, B., 2007. Ciphertext-policy
in: Proceedings of the 24th ACM SIGKDD international conference attribute-based encryption, in: 2007 IEEE symposium on security
on knowledge discovery & data mining, pp. 2867–2867. and privacy (SP’07), IEEE. pp. 321–334.
[2] Al Jawaheri, H., Al Sabah, M., Boshmaf, Y., Erbad, A., 2020. [22] Bhargavan, K., Delignat-Lavaud, A., Fournet, C., Gollamudi, A.,
Deanonymizing tor hidden service users through bitcoin transactions Gonthier, G., Kobeissi, N., Rastogi, A., Sibut-Pinote, T., Swamy, N.,
analysis. Computers & Security 89, 101684. Zanella-Béguelin, S., 2016. Formal verification of smart contracts:
[3] Alberto Torres, W.A., Steinfeld, R., Sakzad, A., Liu, J.K., Kuchta, Short paper, in: Proceedings of the 2016 ACM Workshop on Pro-
V., Bhattacharjee, N., Au, M.H., Cheng, J., 2018. Post-quantum gramming Languages and Analysis for Security, ACM. pp. 91–96.
one-time linkable ring signature and application to ring confidential [23] Biryukov, A., Tikhomirov, S., 2019a. Deanonymization and link-
transactions in blockchain (lattice ringct v1. 0), in: Information ability of cryptocurrency transactions based on network analysis,
Security and Privacy: 23rd Australasian Conference, ACISP 2018, in: 2019 IEEE European symposium on security and privacy (Eu-
Wollongong, NSW, Australia, July 11-13, 2018, Proceedings 23, roS&P), IEEE. pp. 172–184.
Springer. pp. 558–576. [24] Biryukov, A., Tikhomirov, S., 2019b. Security and privacy of mobile
[4] Androulaki, E., Karame, G.O., Roeschlin, M., Scherer, T., Cap- wallet users in bitcoin, dash, monero, and zcash. Pervasive and
kun, S., 2013. Evaluating user privacy in bitcoin, in: Financial Mobile Computing 59, 101030.
Cryptography and Data Security: 17th International Conference, FC [25] Biryukov, A., Tikhomirov, S., 2019c. Transaction clustering using
network traffic analysis for bitcoin and derived blockchains, in: IEEE

Jamwal et al.: Preprint submitted to Elsevier Page 20 of 26

 Survey on Ethereum Pseudonymity

INFOCOM 2019-IEEE Conference on Computer Communications Tutorials 20, 3416–3452.

Workshops (INFOCOM WKSHPS), IEEE. pp. 204–209. [47] Courtois, N.T., Mercer, R., 2017. Stealth address and key
[26] Bitansky, N., Canetti, R., Chiesa, A., Tromer, E., 2013. Recursive management techniques in blockchain systems, in: ICISSP 2017-
composition and bootstrapping for snarks and proof-carrying data, Proceedings of the 3rd International Conference on Information
in: Proceedings of the forty-fifth annual ACM symposium on Theory Systems Security and Privacy, pp. 559–566.
of computing, pp. 111–120. [48] Cramer, R., Damgård, I., Maurer, U., 2000. General secure multi-
[27] Blum, M., De Santis, A., Micali, S., Persiano, G., 1991. Noninterac- party computation from any linear secret-sharing scheme, in: In-
tive zero-knowledge. SIAM Journal on Computing 20, 1084–1118. ternational Conference on the Theory and Applications of Crypto-
[28] Bobba, R., Khurana, H., Prabhakaran, M., 2009. Attribute-sets: A graphic Techniques, Springer. pp. 316–334.
practically motivated enhancement to attribute-based encryption, in: [49] Dannen, C., 2017. Introducing Ethereum and solidity. volume 1.
Computer Security–ESORICS 2009: 14th European Symposium on Springer.
Research in Computer Security, Saint-Malo, France, September 21- [50] De Angelis, S., Aniello, L., Baldoni, R., Lombardi, F., Margheri,
23, 2009. Proceedings 14, Springer. pp. 587–604. A., Sassone, V., et al., 2018. Pbft vs proof-of-authority: Applying
[29] Bonneau, J., Narayanan, A., Miller, A., Clark, J., Kroll, J.A., Felten, the cap theorem to permissioned blockchain, in: CEUR workshop
E.W., 2014. Mixcoin: Anonymity for bitcoin with accountable proceedings, CEUR-WS. pp. 1–11.
mixes, in: International Conference on Financial Cryptography and [51] De Filippi, P., Wright, A., 2018. Blockchain and the law: The rule
Data Security, Springer. pp. 486–504. of code. Harvard University Press.
[30] Boyle, E., Gilboa, N., Ishai, Y., Nof, A., 2019. Practical fully secure [52] Deuber, D., Schröder, D., 2021. Coinjoin in the wild: An empirical
three-party computation via sublinear distributed zero-knowledge analysis in dash, in: Computer Security–ESORICS 2021: 26th Eu-
proofs, in: Proceedings of the 2019 ACM SIGSAC Conference on ropean Symposium on Research in Computer Security, Darmstadt,
Computer and Communications Security, pp. 869–886. Germany, October 4–8, 2021, Proceedings, Part II 26, Springer. pp.
[31] Bünz, B., Agrawal, S., Zamani, M., Boneh, D., 2020. Zether: To- 461–480.
wards privacy in a smart contract world, in: International Conference [53] Dijk, M.v., Gentry, C., Halevi, S., Vaikuntanathan, V., 2010. Fully
on Financial Cryptography and Data Security, Springer. pp. 423– homomorphic encryption over the integers, in: Annual international
443. conference on the theory and applications of cryptographic tech-
[32] Buterin, V., 2014. A next-generation smart contract and decentral- niques, Springer. pp. 24–43.
ized application platform. white paper . [54] Dorri, A., Steger, M., Kanhere, S.S., Jurdak, R., 2017. Blockchain:
[33] Buterin, V., 2016. Ethereum: platform review. Opportunities and A distributed solution to automotive security and privacy. IEEE
Challenges for Private and Consortium Blockchains 45. Communications Magazine 55, 119–125.
[34] Čapko, D., Vukmirović, S., Nedić, N., 2022. State of the art of zero- [55] Du, W., Atallah, M.J., 2001. Secure multi-party computation
knowledge proofs in blockchain, in: 2022 30th Telecommunications problems and their applications: a review and open problems, in:
Forum (TELFOR), IEEE. pp. 1–4. Proceedings of the 2001 workshop on New security paradigms, pp.
[35] Chan, W., Olmsted, A., 2017. Ethereum transaction graph analysis, 13–22.
in: 2017 12th international conference for internet technology and [56] Duffield, E., Diaz, D., 2015. Dash: A privacycentric cryptocurrency.
secured transactions (ICITST), IEEE. pp. 498–500. [57] Dwork, C., 2006. Differential privacy, in: International colloquium
[36] Chase, M., 2007. Multi-authority attribute based encryption, in: on automata, languages, and programming, Springer. pp. 1–12.
Theory of Cryptography: 4th Theory of Cryptography Conference, [58] Dwork, C., 2008. Differential privacy: A survey of results, in:
TCC 2007, Amsterdam, The Netherlands, February 21-24, 2007. International conference on theory and applications of models of
Proceedings 4, Springer. pp. 515–534. computation, Springer. pp. 1–19.
[37] Chase, M., Chow, S.S., 2009. Improving privacy and security in [59] Dwork, C., Roth, A., 2014. The algorithmic foundations of differen-
multi-authority attribute-based encryption, in: Proceedings of the tial privacy. Found. Trends Theor. Comput. Sci. 9, 211–407. URL:
16th ACM conference on Computer and communications security, https://doi.org/10.1561/0400000042, doi:10.1561/0400000042.
pp. 121–130. [60] Dwork, C., Roth, A., et al., 2014. The algorithmic foundations
[38] Chaum, D.L., 1981. Untraceable electronic mail, return addresses, of differential privacy. Foundations and Trends® in Theoretical
and digital pseudonyms. Communications of the ACM 24, 84–90. Computer Science 9, 211–407.
[39] Chawathe, S.S., 2019. Clustering blockchain data. Clustering Meth- [61] Dziembowski, S., Faust, S., Hostáková, K., 2018. General state chan-
ods for Big Data Analytics: Techniques, Toolboxes and Applications nel networks, in: Proceedings of the 2018 ACM SIGSAC Conference
, 43–72. on Computer and Communications Security, pp. 949–966.
[40] Chen, C., Asoni, D.E., Barrera, D., Danezis, G., Perrig, A., 2015. [62] Ermilov, D., Panov, M., Yanovich, Y., 2017a. Automatic bitcoin
Hornet: High-speed onion routing at the network layer, in: Pro- address clustering, in: 2017 16th IEEE International Conference
ceedings of the 22nd ACM SIGSAC Conference on Computer and on Machine Learning and Applications (ICMLA), pp. 461–466.
Communications Security, pp. 1441–1454. doi:10.1109/ICMLA.2017.0-118.
[41] Chen, T., Li, Z., Luo, X., 2017. Understanding ethereum via graph [63] Ermilov, D., Panov, M., Yanovich, Y., 2017b. Automatic bitcoin
analysis, in: 2018 IEEE European Symposium on Security and address clustering, in: 2017 16th IEEE International Conference on
Privacy (EuroS&P), IEEE. pp. 137–152. Machine Learning and Applications (ICMLA), IEEE. pp. 461–466.
[42] Chen, T., Li, Z., Zhu, Y., Chen, J., Luo, X., Lui, J.C.S., Lin, X., [64] European Parliament and EU Council, 2016. Regulation (eu)
Zhang, X., 2020. Understanding ethereum via graph analysis. ACM 2016/679 of the european parliament and of the council. Official
Transactions on Internet Technology (TOIT) 20, 1–32. Journal of the European Union 679, 2016. URL: https://eur-lex.
[43] Christensen, S., 2018. A comparative study of privacy-preserving europa.eu/legal-content/EN/TXT/?uri=CELEX:32016R0679.
cryptocurrencies: Monero and zcash. School of Computer Science [65] Eyal, I., Sirer, E.G., 2018. Majority is not enough: Bitcoin mining
University of Birmingham . is vulnerable. Association for Computing Machinery 61, 95–102.
[44] Christidis, K., Devetsikiotis, M., 2016. Blockchains and smart URL: https://doi.org/10.1145/3212998, doi:10.1145/3212998.
contracts for the internet of things. Ieee Access 4, 2292–2303. [66] Fang, W., Chen, W., Zhang, W., Pei, J., Gao, W., Wang, G.,
[45] Clack, C.D., Bakshi, V.A., Braine, L., 2016. Smart contract tem- 2020. Digital signature scheme for information non-repudiation in
plates: foundations, design landscape and research directions. arXiv blockchain: a state of the art review. EURASIP Journal on Wireless
preprint arXiv:1608.00771 . Communications and Networking 2020, 1–15.
[46] Conti, M., Kumar, E.S., Lal, C., Ruj, S., 2018. A survey on security
and privacy issues of bitcoin. IEEE Communications Surveys &

Jamwal et al.: Preprint submitted to Elsevier Page 21 of 26

 Survey on Ethereum Pseudonymity

[67] Feige, U., Lapidot, D., Shamir, A., 1990. Multiple non-interactive parallel and distributed systems 23, 2150–2162.
zero knowledge proofs based on a single random string, in: Proceed- [88] Han, S., Park, S., 2022. A gap between blockchain and general data
ings [1990] 31st Annual Symposium on Foundations of Computer protection regulation: A systematic review. IEEE Access .
Science, IEEE. pp. 308–317. [89] Harrigan, M., Fretter, C., 2016. The unreasonable effectiveness
[68] Fekete, D.L., Kiss, A., 2024. Trust-minimized optimistic cross- of address clustering, in: 2016 Intl IEEE Conferences on Ubiq-
rollup arbitrary message bridge. Journal of Network and Computer uitous Intelligence & Computing, Advanced and Trusted Com-
Applications 221, 103771. puting, Scalable Computing and Communications, Cloud and Big
[69] Feng, Q., He, D., Zeadally, S., Khan, M.K., Kumar, N., 2019. A Data Computing, Internet of People, and Smart World Congress
survey on privacy protection in blockchain system. Journal of (UIC/ATC/ScalCom/CBDCom/IoP/SmartWorld), IEEE. pp. 368–
Network and Computer Applications 126, 45–58. 373.
[70] Fiege, U., Fiat, A., Shamir, A., 1987. Zero knowledge proofs of [90] Hassan, M.U., Rehmani, M.H., Chen, J., 2020. Differential privacy
identity, in: Proceedings of the nineteenth annual ACM symposium in blockchain technology: A futuristic approach. Journal of Parallel
on Theory of computing, pp. 210–217. and Distributed Computing 145, 50–74.
[71] Finck, M., 2018. Blockchains and data protection in the european [91] Hasselgren, A., Wan, P.K., Horn, M., Kralevska, K., Gligoroski, D.,
union. Eur. Data Prot. L. Rev. 4, 17. Faxvaag, A., 2020. Gdpr compliance for blockchain applications in
[72] Fleder, M., Kester, M.S., Pillai, S., 2015. Bitcoin transaction graph healthcare. Computer Science & Information Technology .
analysis. arXiv preprint arXiv:1502.01657 . [92] Henry, R., Herzberg, A., Kate, A., 2018. Blockchain access privacy:
[73] Gai, K., Wu, Y., Zhu, L., Zhang, Z., Qiu, M., 2019. Differential Challenges and directions. IEEE Security & Privacy 16, 38–45.
privacy-based blockchain for industrial internet-of-things. IEEE [93] Hopwood, D., Bowe, S., Hornby, T., Wilcox, N., 2016. Zcash
Transactions on Industrial Informatics 16, 4156–4165. protocol specification. GitHub: San Francisco, CA, USA , 1.
[74] Gaihre, A., Luo, Y., Liu, H., 2018. Do bitcoin users really care [94] Humbeeck, A.V., et al., 2019. The blockchain-gdpr paradox. Journal
about anonymity? an analysis of the bitcoin transaction graph, in: of Data Protection & Privacy 2, 208–212.
2018 ieee international conference on big data (big data), IEEE. pp. [95] Hur, J., 2011. Improving security and efficiency in attribute-based
1198–1207. data sharing. IEEE transactions on knowledge and data engineering
[75] Gaihre, A., Pandey, S., Liu, H., 2019. Deanonymizing cryptocur- 25, 2271–2282.
rency with graph learning: The promises and challenges, in: 2019 [96] Javed, L., Anjum, A., Yakubu, B.M., Iqbal, M., Moqurrab, S.A.,
IEEE Conference on Communications and Network Security (CNS), Srivastava, G., 2023. Sharechain: Blockchain-enabled model for
IEEE. pp. 1–3. sharing patient data using federated learning and differential privacy.
[76] Gamage, C., Gras, B., Crispo, B., Tanenbaum, A.S., 2006. An Expert Systems 40, e13131.
identity-based ring signature scheme with enhanced privacy, in: [97] Jia, B., Zhang, X., Liu, J., Zhang, Y., Huang, K., Liang, Y., 2021.
2006 Securecomm and Workshops, IEEE. pp. 1–5. Blockchain-enabled federated learning data protection aggregation
[77] Gao, Y., Shi, J., Wang, X., Shi, R., Yin, Z., Yang, Y., 2021. Practical scheme with differential privacy and homomorphic encryption in
deanonymization attack in ethereum based on p2p network anal- iiot. IEEE Transactions on Industrial Informatics 18, 4049–4058.
ysis, in: 2021 IEEE Intl Conf on Parallel & Distributed Process- [98] Jiang, Y., Xu, X., Xiao, F., 2022. Attribute-based encryption with
ing with Applications, Big Data & Cloud Computing, Sustainable blockchain protection scheme for electronic health records. IEEE
Computing & Communications, Social Computing & Networking Transactions on Network and Service Management 19, 3884–3895.
(ISPA/BDCloud/SocialCom/SustainCom), IEEE. pp. 1402–1409. [99] Johnson, D., Menezes, A., Vanstone, S., 2001. The elliptic curve dig-
[78] Garfinkel, S.L., Abowd, J.M., Powazek, S., 2018. Issues encountered ital signature algorithm (ecdsa). International journal of information
deploying differential privacy, in: Proceedings of the 2018 Workshop security 1, 36–63.
on Privacy in the Electronic Society, pp. 133–137. [100] Kabla, A.H.H., Anbar, M., Manickam, S., Al-Amiedy, T.A., Cruspe,
[79] Gentry, C., 2009. A fully homomorphic encryption scheme. Stan- P.B., Al-Ani, A.K., Karuppayah, S., 2022. Applicability of intrusion
ford university. detection system on ethereum attacks: A comprehensive review.
[80] Godyn, M., Kedziora, M., Ren, Y., Liu, Y., Song, H., 2022. Analysis IEEE Access 10, 71632–71655.
of solutions for a blockchain compliance with gdpr. Scientific [101] Kalra, S., Goel, S., Dhawan, M., Sharma, S., 2018. Zeus: Analyzing
Reports 12, 15021. doi:10.1038/s41598-022-19341-y. safety of smart contracts, in: Network and Distributed Systems
[81] Goldfeder, S., Kalodner, H., Reisman, D., Narayanan, A., 2018. Security (NDSS) Symposium, pp. 1–12.
When the cookie meets the blockchain: Privacy risks of web pay- [102] Kang, C., Lee, C., Ko, K., Woo, J., Hong, J.W.K., 2020. De-
ments via cryptocurrencies. Proceedings on Privacy Enhancing anonymization of the bitcoin network using address clustering, in:
Technologies 4, 179–199. Blockchain and Trustworthy Systems: Second International Con-
[82] Goldreich, O., Oren, Y., 1994. Definitions and properties of zero- ference, BlockSys 2020, Dali, China, August 6–7, 2020, Revised
knowledge proof systems. Journal of Cryptology 7, 1–32. Selected Papers 2, Springer. pp. 489–501.
[83] Goldwasser, S., Micali, S., Rackoff, C., 2019. The knowledge [103] Kappos, G., Yousaf, H., Maller, M., Meiklejohn, S., 2018. An
complexity of interactive proof-systems, in: Providing Sound Foun- empirical analysis of anonymity in zcash, in: 27th USENIX Security
dations for Cryptography: On the Work of Shafi Goldwasser and Symposium (USENIX Security 18), pp. 463–477.
Silvio Micali, pp. 203–225. [104] Kelen, D.M., Seres, I.A., Béres, F., Benczúr, A.A., 2023. In-
[84] Goyal, V., Pandey, O., Sahai, A., Waters, B., 2006. Attribute- tegrated onion routing for peer-to-peer validator privacy in the
based encryption for fine-grained access control of encrypted data, ethereum network. URL: https://info.ilab.sztaki.hu/~kdomokos/
in: Proceedings of the 13th ACM conference on Computer and OnionRoutingP2PEthereumPrivacy.pdf.
communications security, pp. 89–98. [105] Kenthapadi, K., Mironov, I., Thakurta, A.G., 2019. Privacy-
[85] Guan, Z., Wan, Z., Yang, Y., Zhou, Y., Huang, B., 2020. Blockmaze: preserving data mining in industry, in: Proceedings of the Twelfth
An efficient privacy-preserving account-model blockchain based on ACM International Conference on Web Search and Data Mining,
zk-snarks. IEEE Transactions on Dependable and Secure Computing pp. 840–841.
19, 1446–1463. [106] Kerber, T., Kiayias, A., Kohlweiss, M., 2021. Kachina – foundations
[86] Halpin, H., Piekarska, M., 2017. Introduction to security and privacy of private smart contracts, in: 2021 IEEE 34th Computer Security
on the blockchain, in: 2017 IEEE European Symposium on Security Foundations Symposium (CSF), pp. 1–16. doi:10.1109/CSF51468.
and Privacy Workshops, IEEE. pp. 1–3. 2021.00002.
[87] Han, J., Susilo, W., Mu, Y., Yan, J., 2012. Privacy-preserving decen- [107] Khan, A., 2022. Graph analysis of the ethereum blockchain data:
tralized key-policy attribute-based encryption. IEEE transactions on A survey of datasets, methods, and future work, in: 2022 IEEE

Jamwal et al.: Preprint submitted to Elsevier Page 22 of 26

 Survey on Ethereum Pseudonymity

International Conference on Blockchain (Blockchain), IEEE. pp. [126] Meiklejohn, S., Mercer, R., 2018. Möbius: Trustless tumbling for
250–257. transaction privacy. Proceedings on Privacy Enhancing Technolo-
[108] Khan, S.N., Loukil, F., Ghedira-Guegan, C., Benkhelifa, E., Bani- gies 2018. doi:10.1515/popets-2018-0015.
Hani, A., 2021. Blockchain smart contracts: Applications, chal- [127] Meiklejohn, S., Pomarole, M., Jordan, G., Levchenko, K., McCoy,
lenges, and future trends. Peer-to-peer Networking and Applications D., Voelker, G.M., Savage, S., 2013. A fistful of bitcoins: character-
14, 2901–2925. izing payments among men with no names, in: Proceedings of the
[109] Khovratovich, D., Vladimirov, M., 2019. Tornado privacy solution: 2013 conference on Internet measurement conference, pp. 127–140.
Cryptographic review. version 1.1. ABDK Consulting, November [128] Merani, M.L., Croce, D., Tinnirello, I., 2021. Rings for privacy: an
29. architecture for large scale privacy-preserving data mining. IEEE
[110] Kılıç, B., Özturan, C., Sen, A., 2022. Parallel analysis of ethereum Transactions on Parallel and Distributed Systems 32, 1340–1352.
blockchain transaction data using cluster computing. Cluster Com- [129] Mercer, R., 2016. Privacy on the blockchain: Unique ring signatures.
puting 25, 1885–1898. arXiv:1612.01188.
[111] Killer, C., Eck, M., Rodrigues, B., von der Assen, J., Staubli, R., [130] Messias, J., Pahari, V., Chandrasekaran, B., Gummadi, K.P.,
Stiller, B., 2022. Provotumn: Decentralized, mix-net-based, and Loiseau, P., 2023. Dissecting bitcoin and ethereum transactions: On
receipt-free voting system, in: 2022 IEEE International Conference the lack of transaction contention and prioritization transparency in
on Blockchain and Cryptocurrency (ICBC), IEEE. pp. 1–9. blockchains. arXiv preprint arXiv:2302.06962 .
[112] Kim, S., Kwon, Y., Cho, S., 2018. A survey of scalability solutions [131] Moralis, . Breaking down eth 2.0 - zk-snarks and zk-
on blockchain, in: 2018 International Conference on Information and rollups. https://academy.moralis.io/blog/breaking-down-eth-2-0-zk-
Communication Technology Convergence (ICTC), IEEE. pp. 1204– snarks-and-zk-rollups. Accessed: 2023-12-01.
1207. [132] Müller, S., Katzenbeisser, S., Eckert, C., 2009. Distributed attribute-
[113] Kosba, A., Miller, A., Shi, E., Wen, Z., Papamanthou, C., based encryption, in: Information Security and Cryptology–ICISC
2016. Hawk: The blockchain model of cryptography and privacy- 2008: 11th International Conference, Seoul, Korea, December 3-5,
preserving smart contracts, in: 2016 IEEE symposium on security 2008, Revised Selected Papers 11, Springer. pp. 20–36.
and privacy (SP), IEEE. pp. 839–858. [133] Naehrig, M., Lauter, K., Vaikuntanathan, V., 2011. Can homo-
[114] Kus, M.C., Levi, A., 2022. Investigation and application of differen- morphic encryption be practical?, in: Proceedings of the 3rd ACM
tial privacy in bitcoin. IEEE Access 10, 25534–25554. doi:10.1109/ workshop on Cloud computing security workshop, pp. 113–124.
ACCESS.2022.3151784. [134] Nakamoto, S., 2008. Bitcoin: A peer-to-peer electronic cash system.
[115] Kushwaha, S.S., Joshi, S., Singh, D., Kaur, M., Lee, H.N., 2022. Decentralized Business Review , 21260.
Systematic review of security vulnerabilities in ethereum blockchain [135] Neiheiser, R., Inácio, G., Rech, L., Montez, C., Matos, M., Ro-
smart contract. IEEE Access 10, 6605–6621. drigues, L., 2023. Practical limitations of ethereum’s layer-2. IEEE
[116] Lapets, A., Albab, K.D., Issa, R., Qin, L., Varia, M., Bestavros, A., Access 11, 8651–8662.
Jansen, F., 2019. Role-based ecosystem for the design, development, [136] Neudecker, T., Andelfinger, P., Hartenstein, H., 2016. Tim-
and deployment of secure multi-party data analytics applications, in: ing analysis for inferring the topology of the bitcoin peer-
2019 IEEE Cybersecurity Development (SecDev), IEEE. pp. 129– to-peer network, in: 2016 Intl IEEE Conferences on Ubiqui-
140. tous Intelligence & Computing, Advanced and Trusted Comput-
[117] Lavaur, T., Lacan, J., Chanel, C.P., 2022. Enabling blockchain ing, Scalable Computing and Communications, Cloud and Big
services for ioe with zk-rollups. Sensors 22, 6493. Data Computing, Internet of People, and Smart World Congress
[118] Li, X., Mei, Y., Gong, J., Xiang, F., Sun, Z., 2020a. A blockchain (UIC/ATC/ScalCom/CBDCom/IoP/SmartWorld), IEEE. pp. 358–
privacy protection scheme based on ring signature. IEEE Access 8, 367.
76765–76772. [137] Nguyen, C.T., Hoang, D.T., Nguyen, D.N., Niyato, D., Nguyen,
[119] Li, Y., Islambekov, U., Akcora, C., Smirnova, E., Gel, Y.R., Kantar- H.T., Dutkiewicz, E., 2019. Proof-of-stake consensus mechanisms
cioglu, M., 2020b. Dissecting ethereum blockchain analytics: What for future blockchain networks: Fundamentals, applications and
we learn from topology and geometry of the ethereum graph?, in: opportunities. IEEE Access 7, 85727–85745. doi:10.1109/ACCESS.
Proceedings of the 2020 SIAM international conference on data 2019.2925010.
mining, SIAM. pp. 523–531. [138] Nikolakis, W., John, L., Krishnan, H., 2018. How blockchain can
[120] Lima, C., 2018. Blockchain gdpr privacy by design. IEEE shape sustainable global value chains: An evidence, verifiability, and
blockchain group 4. enforceability (eve) framework. Sustainability 10, 3926.
[121] Liu, Jiangyan, Cai, Zhiyuan, Liu, Dunnan, Jin, Tianyi, 2021. [139] Niranjanamurthy, M., Nithya, B., Jagannatha, S., 2019. Analysis of
Research on distributed energy transaction technology based on blockchain technology: pros, cons and swot. Cluster Computing 22,
blockchain. E3S Web Conf. 236, 02011. URL: https://doi.org/ 14743–14757.
10.1051/e3sconf/202123602011, doi:10.1051/e3sconf/202123602011. [140] Noether, S., Mackenzie, A., et al., 2016. Ring confidential transac-
[122] Lotti, L., 2016. Contemporary art, capitalization and the blockchain: tions. Ledger 1, 1–18.
On the autonomy and automation of art’s value. Finance and Society [141] Sáez de Ocáriz Borde, H., 2022. An overview of trees in blockchain
2, 96–110. technology: Merkle trees and merkle patricia tries. Department of
[123] Lu, H., Yu, R., Zhu, Y., He, X., Liang, K., Chu, W.C.C., Engineering, University of Cambridge .
2022. Policy-driven data sharing over attribute-based [142] Omar, I.A., Jayaraman, R., Salah, K., Simsekler, M.C.E., Yaqoob,
encryption supporting dual membership. Journal of I., Ellahham, S., 2020. Ensuring protocol compliance and data
Systems and Software 188, 111271. URL: https://www. transparency in clinical trials using blockchain smart contracts.
sciencedirect.com/science/article/pii/S0164121222000346, BMC Medical Research Methodology 20, 1–17.
doi:https://doi.org/10.1016/j.jss.2022.111271. [143] Ouaddah, A., Elkalam, A.A., Ouahman, A.A., 2017. Towards a
[124] Maurer, F.K., Neudecker, T., Florian, M., 2017. Anonymous novel privacy-preserving access control model based on blockchain
coinjoin transactions with arbitrary values, in: 2017 ieee trust- technology in iot, in: Rocha, Á., Serrhini, M., Felgueiras, C. (Eds.),
com/bigdatase/icess, IEEE. pp. 522–529. Europe and MENA Cooperation Advances in Information and Com-
[125] McSherry, F., Talwar, K., 2007. Mechanism design via differential munication Technologies, Springer International Publishing, Cham.
privacy, in: 48th Annual IEEE Symposium on Foundations of Com- pp. 523–533.
puter Science (FOCS’07), IEEE. pp. 94–103. [144] Panait, A.E., Olimid, R.F., 2021. On using zk-snarks and zk-
starks in blockchain-based identity management, in: Innovative Se-
curity Solutions for Information Technology and Communications:

Jamwal et al.: Preprint submitted to Elsevier Page 23 of 26

 Survey on Ethereum Pseudonymity

13th International Conference, SecITC 2020, Bucharest, Romania, through blockchain technology. Futures 122, 102611.
November 19–20, 2020, Revised Selected Papers 13, Springer. pp. [166] Seres, I.A., Nagy, D.A., Buckland, C., Burcsi, P., 2019. Mixeth:
130–145. efficient, trustless coin mixing service for ethereum. Cryptology
[145] Park, Y.H., Kim, Y., Shim, J., 2021. Blockchain-based privacy- ePrint Archive doi:10.4230/OASIcs.Tokenomics.2019.13.
preserving system for genomic data management using local dif- [167] Shi, Y., Cao, K., Yao, J., Ge, X., 2023. A semi-homomorphic privacy
ferential privacy. Electronics 10. URL: https://www.mdpi.com/ computing solution based on sm2 and blockchain, in: International
2079-9292/10/23/3019, doi:10.3390/electronics10233019. Conference on Cryptography, Network Security, and Communica-
[146] Partala, J., Nguyen, T.H., Pirttikangas, S., 2020. Non-interactive tion Technology (CNSCT 2023), SPIE. pp. 29–38.
zero-knowledge for blockchain: A survey. IEEE Access 8, 227945– [168] Shojaeinasab, A., Motamed, A.P., Bahrak, B., 2023. Mixing detec-
227961. tion on bitcoin transactions using statistical patterns. IET Blockchain
[147] Pertsev, A., Semenov, R., Storm, R., 2019. Tornado cash privacy 3, 136–148.
solution version 1.4. Tornado cash privacy solution version 1. [169] Singh, K., Heulot, N., Hamida, E.B., 2018. Towards anonymous,
[148] Petrov, S., Kendzierskyj, S., Jahankhani, H., 2020. Protecting unlinkable, and confidential transactions in blockchain, in: 2018
privacy and security using tor and blockchain and de-anonymization IEEE International Conference on Internet of Things (iThings) and
risks. Cyber Defence in the Age of AI, Smart Societies and IEEE Green Computing and Communications (GreenCom) and
Augmented Humanity , 199–232. IEEE Cyber, Physical and Social Computing (CPSCom) and IEEE
[149] Pinto, A.M., 2020. An introduction to the use of zk-snarks in Smart Data (SmartData), IEEE. pp. 1642–1649.
blockchains, in: Mathematical Research for Blockchain Economy. [170] Singh, N.P., 2021. Regulating Cross Border Data Flows: An Assess-
Springer, pp. 233–249. ment of India’s Data Localisation Framework. Ph.D. thesis. National
[150] Politou, E., Casino, F., Alepis, E., Patsakis, C., 2019. Blockchain Law School of India University, Bangalore.
mutability: Challenges and proposed solutions. IEEE Transactions [171] Stevens, R., 2022. What are ethereum rollups? a scaling solution
on Emerging Topics in Computing 9, 1972–1986. to cut transaction costs. URL: https://decrypt.co/resources/
[151] Poon, J., Buterin, V., 2017. Plasma: Scalable autonomous smart what-are-ethereum-rollups-scaling-solution-cut-transaction-costs.
contracts. White paper , 1–47. [172] Suma, V., 2019. Security and privacy mechanism using blockchain.
[152] Rackoff, C., Simon, D.R., 1991. Non-interactive zero-knowledge Journal of Ubiquitous Computing and Communication Technologies
proof of knowledge and chosen ciphertext attack, in: Annual inter- (UCCT) 1, 45–54.
national cryptology conference, Springer. pp. 433–444. [173] Sun, H., Ruan, N., Liu, H., 2019. Ethereum analysis via node
[153] Rahulamathavan, Y., Phan, R.C.W., Rajarajan, M., Misra, S., Kon- clustering, in: Network and System Security: 13th International
doz, A., 2017. Privacy-preserving blockchain based iot ecosystem Conference, NSS 2019, Sapporo, Japan, December 15–18, 2019,
using attribute-based encryption, in: 2017 IEEE International Con- Proceedings 13, Springer. pp. 114–129.
ference on Advanced Networks and Telecommunications Systems [174] Sun, S.F., Au, M.H., Liu, J.K., Yuen, T.H., 2017. Ringct 2.0: A
(ANTS), IEEE. pp. 1–6. compact accumulator-based (linkable ring signature) protocol for
[154] Rao, U.P., Shukla, P.K., Trivedi, C., Gupta, S., Shibeshi, Z.S., 2021. blockchain cryptocurrency monero, in: European Symposium on
Blockchain for Information Security and Privacy. CRC Press. Research in Computer Security, Springer. pp. 456–474.
[155] Reynolds, P., Irwin, A.S., 2017. Tracking digital footprints: [175] Sun, X., Yu, F.R., Zhang, P., Sun, Z., Xie, W., Peng, X., 2021. A
anonymity within the bitcoin system. Journal of Money Laundering survey on zero-knowledge proof in blockchain. IEEE network 35,
Control 20, 172–189. 198–205.
[156] Rivest, R.L., Shamir, A., Tauman, Y., 2006. How to leak a secret: [176] Suripeddi, M.K.S., Purandare, P., 2021. Blockchain and gdpr – a
Theory and applications of ring signatures. Theoretical Computer study on compatibility issues of the distributed ledger technology
Science: Essays in Memory of Shimon Even , 164–186. with gdpr data processing. Journal of Physics: Conference Series
[157] Ron, D., Shamir, A., 2013a. Quantitative analysis of the full bitcoin 1964. URL: https://api.semanticscholar.org/CorpusID:236205971.
transaction graph, in: Sadeghi, A.R. (Ed.), Financial Cryptography [177] Szabo, N., 1997. Formalizing and securing relationships on public
and Data Security, Springer Berlin Heidelberg, Berlin, Heidelberg. networks. First monday .
pp. 6–24. doi:10.1007/978-3-642-39884-1\_2. [178] Tang, Y., Xu, C., Zhang, C., Wu, Y., Zhu, L., 2021. Analysis of
[158] Ron, D., Shamir, A., 2013b. Quantitative analysis of the full bitcoin address linkability in tornado cash on ethereum. ieee , 39–50.
transaction graph, in: Financial Cryptography and Data Security: [179] Team, E., 2024. Republic protocol (ren) review: Complete be-
17th International Conference, FC 2013, Okinawa, Japan, April 1-5, ginners guide to ren. URL: https://www.coinbureau.com/review/
2013, Revised Selected Papers 17, Springer. pp. 6–24. republic-protocol-ren/.
[159] Sabry, F., Labda, W., Erbad, A., Malluhi, Q., 2020. Cryptocurrencies [180] Thibault, L.T., Sarry, T., Hafid, A.S., 2022. Blockchain scaling using
and artificial intelligence: Challenges and opportunities. IEEE rollups: A comprehensive survey. IEEE Access .
Access 8, 175840–175858. doi:10.1109/ACCESS.2020.3025211. [181] Thyagarajan, S.A., Malavolta, G., Moreno-Sanchez, P., 2022.
[160] Sahai, A., Waters, B., 2005. Fuzzy identity-based encryption, in: Universal atomic swaps: Secure exchange of coins across all
Annual international conference on the theory and applications of blockchains, in: 2022 IEEE Symposium on Security and Privacy
cryptographic techniques, Springer. pp. 457–473. (SP), IEEE. pp. 1299–1316.
[161] Sahu, N., Gajera, M., Chaudhary, A., 2023. zkfi: Privacy-preserving [182] Tovanich, N., Cazabet, R., 2023. Fingerprinting bitcoin entities
and regulation compliant transactions using zero knowledge proofs. using money flow representation learning. Applied Network Science
arXiv preprint arXiv:2307.00521 . 8, 63.
[162] Satybaldy, A., Nowostawski, M., 2020. Review of techniques for [183] Truong, N.B., Sun, K., Lee, G.M., Guo, Y., 2020. Gdpr-
privacy-preserving blockchain systems, in: Proceedings of the 2nd compliant personal data management: A blockchain-based solu-
ACM International Symposium on Blockchain and Secure Critical tion. IEEE Transactions on Information Forensics and Security
Infrastructure, pp. 1–9. 15, 1746–1761. URL: https://doi.org/10.1109/TIFS.2019.2948287,
[163] Schaffner, T., 2021. Scaling public blockchains: A comprehensive doi:10.1109/TIFS.2019.2948287.
analysis of optimistic and zero-knowledge rollups. Master Thesis: [184] Valdeolmillos, D., Mezquita, Y., González-Briones, A., Prieto, J.,
University of Basel . Corchado, J.M., 2020. Blockchain technology: a review of the cur-
[164] Schmitz, A., Rule, C., 2019. Online dispute resolution for smart rent challenges of cryptocurrency, in: Blockchain and Applications:
contracts. J. Disp. Resol. , 103. International Congress, Springer. pp. 153–160.
[165] Schulz, K.A., Gstrein, O.J., Zwitter, A.J., 2020. Exploring the gov- [185] Van Saberhagen, N., 2013. Cryptonote v2.0. URL: https://
ernance and implementation of sustainable development initiatives cryptonote.org/whitepaper.pdf.

Jamwal et al.: Preprint submitted to Elsevier Page 24 of 26

 Survey on Ethereum Pseudonymity

[186] Varghese, H.M., Nagoree, D.A., Jayapandian, N., et al., 2021. Cryp- [204] Yakovenko, A., 2018. Solana: A new architecture for a high
tocurrency security and privacy issues: A research perspective, in: performance blockchain v0. 8.13. Whitepaper .
2021 6th International Conference on Communication and Electron- [205] Yan, Y., Shao, G., Song, D., Song, M., Jin, Y., 2023. He-dksap:
ics Systems (ICCES), IEEE. pp. 902–907. Privacy-preserving stealth address protocol via additively homomor-
[187] Victor, F., 2020. Address clustering heuristics for ethereum, in: phic encryption. arXiv:2312.10698.
Financial Cryptography and Data Security: 24th International Con- [206] Yang, K., Jia, X., 2012. Attributed-based access control for multi-
ference, FC 2020, Kota Kinabalu, Malaysia, February 10–14, 2020 authority systems in cloud storage, in: 2012 IEEE 32nd International
Revised Selected Papers 24, Springer. pp. 617–633. Conference on Distributed Computing Systems, IEEE. pp. 536–545.
[188] Vo, H.T., Kundu, A., Mohania, M.K., 2018. Research directions in [207] Yao, A.C., 1982. Protocols for secure computations, in: 23rd annual
blockchain data management and analytics., in: EDBT, pp. 445–448. symposium on foundations of computer science (sfcs 1982), IEEE.
[189] Wang, B., Li, H., Guo, Y., Wang, J., 2023. Ppflhe: A privacy- pp. 160–164.
preserving federated learning scheme with homomorphic encryption [208] Ye, C., Li, G., Cai, H., Gu, Y., Fukuda, A., 2018. Analysis
for healthcare data. Applied Soft Computing 146, 110677. of security in blockchain: Case study in 51%-attack detecting, in:
[190] Wang, G., Liu, Q., Wu, J., 2010. Hierarchical attribute-based 2018 5th International conference on dependable systems and their
encryption for fine-grained access control in cloud storage services, applications (DSA), IEEE. pp. 15–24.
in: Proceedings of the 17th ACM conference on Computer and [209] Zhang, J., Bai, W., Jiang, Z., 2020a. On the security of a practical
communications security, pp. 735–737. constant-size ring signature scheme. Int. J. Netw. Secur. 22, 392–
[191] Wang, Q., Zhang, Y., Bao, Z., Shi, W., Lei, H., Liu, H., Chen, 396.
B., 2022. Sortee: Service-oriented routing for payment channel [210] Zhang, M., Zhang, X., Zhang, Y., Lin, Z., 2020b. {TXSPECTOR}:
networks with scalability and privacy protection. IEEE Transactions Uncovering attacks in ethereum from transactions, in: 29th USENIX
on Network and Service Management 19, 3764–3780. doi:10.1109/ Security Symposium (USENIX Security 20), pp. 2775–2792.
TNSM.2022.3213559. [211] Zhang, R., Xue, R., Liu, L., 2019. Security and privacy on
[192] Wang, W., Hoang, D.T., Hu, P., Xiong, Z., Niyato, D., Wang, P., blockchain. ACM Computing Surveys (CSUR) 52, 1–34.
Wen, Y., Kim, D.I., 2019. A survey on consensus mechanisms and [212] Zhang, Y., Deng, R.H., Xu, S., Sun, J., Li, Q., Zheng, D., 2020c.
mining strategy management in blockchain networks. Ieee Access Attribute-based encryption for cloud computing access control: A
7, 22328–22370. survey. ACM Computing Surveys (CSUR) 53, 1–41.
[193] Wang, Y., Tan, M., 2023. Defense against sybil attack in blockchain [213] Zhao, C., Zhao, S., Zhao, M., Chen, Z., Gao, C.Z., Li, H., Tan,
based on improved consensus algorithm, in: 2023 IEEE International Y.a., 2019. Secure multi-party computation: theory, practice and
Conference on Control, Electronics and Computer Technology (IC- applications. Information Sciences 476, 357–372.
CECT), pp. 986–989. doi:10.1109/ICCECT57938.2023.10140278. [214] Zhao, H., Bai, X., Zheng, S., Wang, L., 2020. Rzcoin: Ethereum-
[194] Werbach, K., 2018. The blockchain and the new architecture of trust. based decentralized payment with optional privacy service. Entropy
MIT Press . 22, 712.
[195] Wilkinson, S., Lowry, J., Boshevski, T., 2014. Metadisk a [215] Zheng, Z., Xie, S., Dai, H., Chen, X., Wang, H., 2017. An
blockchain-based decentralized file storage application. Storj Labs overview of blockchain technology: Architecture, consensus, and
Inc., Technical Report, hal 1. future trends, in: 2017 IEEE international congress on big data
[196] Wind, P., 2021. Vitalik buterin is optimistic on rollups and other (BigData congress), Ieee. pp. 557–564.
layer 2 solutions to bring sufficient scaling to ethereum. URL: [216] Zhong, H., Sang, Y., Zhang, Y., Xi, Z., 2020. Secure multi-party
https://shorturl.at/oFGK8. computation on blockchain: An overview, in: Parallel Architec-
[197] Wohrer, M., Zdun, U., 2018. Smart contracts: security patterns in the tures, Algorithms and Programming: 10th International Symposium,
ethereum ecosystem and solidity, in: 2018 International Workshop PAAP 2019, Guangzhou, China, December 12–14, 2019, Revised
on Blockchain Oriented Software Engineering (IWBOSE), IEEE. Selected Papers 10, Springer. pp. 452–460.
pp. 2–8. [217] Zhou, J., Lam, K.Y., 1999. Securing digital signatures for non-
[198] Wu, A., Zhang, Y., Zheng, X., Guo, R., Zhao, Q., Zheng, D., repudiation. Computer Communications 22, 710–716.
2019. Efficient and privacy-preserving traceable attribute-based [218] Zou, W., Lo, D., Kochhar, P.S., Le, X.B.D., Xia, X., Feng, Y.,
encryption in blockchain. Annals of Telecommunications 74, 401– Chen, Z., Xu, B., 2019. Smart contract development: Challenges
411. and opportunities. IEEE Transactions on Software Engineering 47,
[199] Wu, J., Liu, J., Chen, W., Huang, H., Zheng, Z., Zhang, Y., 2021a. 2084–2106.
Detecting mixing services via mining bitcoin transaction network [219] Zwitter, A., Hazenberg, J., 2020. Decentralized network governance:
with hybrid motifs. IEEE Transactions on Systems, Man, and blockchain technology and the future of regulation. Frontiers in
Cybernetics: Systems 52, 2237–2249. Blockchain 3, 12.
[200] Wu, J., Liu, J., Zhao, Y., Zheng, Z., 2021b. Analysis of cryptocur- [220] Zyskind, G., Nathan, O., Pentland, A., 2015. Enigma: Decentralized
rency transactions from a network perspective: An overview. Journal computation platform with guaranteed privacy. arXiv preprint
of Network and Computer Applications 190, 103139. arXiv:1506.03471 .
[201] Wu, L., Hu, Y., Zhou, Y., Wang, H., Luo, X., Wang, Z., Zhang, F.,
Ren, K., 2021c. Towards understanding and demystifying bitcoin
mixing services, in: Proceedings of the Web Conference 2021, pp.
33–44.
[202] Xu, Z., Sun, Q., Han, H., Dong, X., Yan, Z., Zheng, Z., Tian,
W., 2022. Bmtac: a decentralized, auditable, time-limited,
multi-authority attribute access control scheme in blockchain en-
vironment, in: 2022 IEEE Smartworld, Ubiquitous Intelligence
& Computing, Scalable Computing & Communications, Digital
Twin, Privacy Computing, Metaverse, Autonomous & Trusted
Vehicles (SmartWorld/UIC/ScalCom/DigitalTwin/PriComp/Meta),
IEEE. pp. 1997–2002.
[203] Xue, L., Yu, Y., Li, Y., Au, M.H., Du, X., Yang, B., 2019. Efficient
attribute-based encryption with attribute revocation for assured data
deletion. Information Sciences 479, 640–650.

Jamwal et al.: Preprint submitted to Elsevier Page 25 of 26

 Survey on Ethereum Pseudonymity

Shivani Jamwal is a M.Sc. Research student at the and Networking (2016-2020), IEEE Journal of Se-
School of Computing Science, the University of lected Areas in Communications, and IEEE Trans-
Glasgow, UK. Prior to her current research de- actions on Machine Learning in Communications
gree, she holds a M.Sc. and B.Tech degrees from Networking.
the University of Glasgow and Amity University
Dubai in 2018 and 2016 respectively. She has
worked as a Graduate Teaching Assistant at the Dr. Nguyen Truong is a Lecturer (i.e., Assistant
University of Glasgow. Her research interests are in Professor) at School of Computing Science, the
the domain of Information Security, Data Privacy, University of Glasgow, UK. Previously, he was a
GDPR, and Blockchain technology. Research Associate at Data Science Institute, De-
partment of Computing, Imperial College London,
UK in 2018-2022. He received his Ph.D, MSc,
José Cano is an Associate Professor in the School and BSc degrees from Liverpool John Moores
of Computing Science at the University of Glas- University, UK, Pohang University of Science and
gow, UK, where he leads the Intelligent Computing Technology, Korea, and Hanoi University of Sci-
Laboratory (gicLAB) and is deputy head of the ence and Technology, Vietnam in 2018, 2013, and
Systems research section. His research interests are 2008, respectively. He was a Software Engineer
in the broad areas of Computer Architecture, Com- at DASAN Networks, a leading company on Net-
puter Systems, Compilers, Machine Learning and working Products and Services in South Korea
Security. José received the MS and PhD degrees in 2012-2015. His research interests are includ-
in Computer Science from Universitat Politècnica ing, but are not limited to, Data Privacy, Security,
de València (UPV), Spain, in 2004 and 2012, re- and Trust, Data Management, Distributed Systems,
spectively. He was a postdoctoral researcher at Uni- Blockchain, and Federated Learning.
versitat Politècnica de Catalunya, Spain, between
2012 and 2013, and at the University of Edinburgh,
UK, between 2014 and 2018. He has authored over
45 refereed publications, and has served as co-
organizer, chair and TPC in numerous conferences
and workshops. He was a member of the IEEE
TPDS Review Board between 2020 and 2021, and
is currently associate editor of ACM TACO. He is
a senior member of IEEE and ACM, and a member
of HiPEAC.

Dr. Gyu Myoung Lee received his BS degree from

Hong Ik University and MS, and PhD degrees
from the Korea Advanced Institute of Science and
Technology (KAIST), Korea, in 1999, 2000 and
2007, respectively. He is a Professor at Depart-
ment of Computer Science, Liverpool John Moores
University, UK. He is also with KAIST as an
adjunct professor. His research interests include
Future Networks, IoT, and multimedia services. He
has actively contributed to standardization in ITU-
T as a Rapporteur, oneM2M and IETF. He is chair
of the ITU-T Focus Group on data processing and
management to support IoT and Smart Cities &
Communities.

Nguyen H. Tran is an Associate Professor at the

School of Computer Science, the University of
Sydney. He was an Assistant Professor with De-
partment of Computer Science and Engineering,
Kyung Hee University, from 2012 to 2017. He
received BS and PhD degrees from HCMC Uni-
versity of Technology and Kyung Hee University
in 2005 and 2011, respectively. His research group
has special interests in Distributed compUting, op-
timizAtion, and machine Learning (DUAL group).
He received several best paper awards, including
IEEE ICC 2016 and ACM MSWiM 2019. He
receives the Korea NRF Funding for Basic Sci-
ence and Research 2016-2023, ARC Discovery
Project 2020-2023, and SOAR award 2022-2023.
He serves as an Editor for several journals such
as IEEE Transactions on Green Communications

Jamwal et al.: Preprint submitted to Elsevier Page 26 of 26