Barriers to blockchain-based decentralised energy trading: a systematic review

ABSTRACT

The increasing adoption of clean energy technologies, including solar and wind generation, demand response, energy efficiency, and energy storage (e.g. batteries and electric vehicles) have led to the evolution of the traditional electricity markets from centralised energy trading systems into Distributed Energy Trading (DET) systems. Consequently, savvy business executives are exploring how blockchain might impact their competitive advantage in the emerging DET markets. Due to its salient features of distributed ledger, consensus mechanisms, cryptography, and smart contracts, blockchain technology is being used to provide decentralised trust, immutability, security and privacy, and transparency in DET system. However, integrating blockchain in DET systems is facing technical, administrative, standardisation and economic barriers. Consequently, we seek to conduct a comprehensive market analysis to identify the specific challenges hindering the integration of blockchain in DET systems. Nonetheless, we noticed that there isn't any evaluation and review framework for conducting a systematic literature review on blockchain-based DET systems. Therefore, in this work we first proposed a conceptual evaluation and review framework for conducting a systematic literature review on blockchain-based DET systems. Then, using the proposed framework, we reviewed the current studies on blockchain-based DET systems to the identify specific challenges hindering the adoption of blockchain and their proposed solutions. Our review found that, although there has been tremendous progress in addressing the technical barriers, the administrative, standardisation and economic barriers have grossly been under reviewed.

KEYWORDS:

• Blockchain challenges
• distributed energy trading
• smart grids
• market models cyber-physical system

1. Introduction

1.1. Overview

Recently, the adoption of rooftop solar panels, windmills, electric vehicle (EVs), and lithium ion batteries, has shown a significant market uptake (Coignard et al. 2018). However, the traditional power grids were not designs to handle the intermittent generation from solar and wind, variable loads from EV and batteries, and reverse energy flow from consumers to the grid. Therefore, as the power grids host more of these Distributed Energy Resources (DERs), the Distribution System Operators (DSOs) face new operational challenges such as increased operating cost, congestion, and variable frequency in the distribution grid (Coignard et al. 2018). Accordingly, the advancement in Information and Communication Technologies (ICT), such as artificial intelligence, bi-directional communication, and Internet-of-Things (IoT), the power grids are becoming 'smart' (Abdella and Shuaib 2018). Figure 1 schematically shows the evolution of the power grid from the traditional hierarchical power grid system in Figure 1(a) to the emerging distributed smart grid system in Figure 1(b). Unlike the traditional power grids, power and information in smart grids can flow in both directions, there is distributed generation, and energy can be stored for a short amount of time.

Figure 1. Transition in electric power grid. (a) Traditional power grid system. (b) Smart grid system.

Most significantly, the evolution of the smart grid has been attributed to the transformation of customers from passive consumers to prosumers. Prosumers are energy consumers who can simultaneously produce and consume electric energy (Ackermann, Andersson, and Söder 2001). Although the prosumers intended to meet their internal demand, their surplus generation could meet their neighbours' demand. Consequently, Governments are providing energy subsidies to incentivise renewable energy generation through market mechanisms such as feed-in tariffs (Ilic et al. 2012). However, feed-in tariffs are slowly being faced-out. As a result, Decentralised Energy Trading (DET) systems such as wholesale energy trading and Peer-to-Peer (P2P) platforms are emerging to support renewable energy generation (Shrestha et al. 2019). In the P2P energy trading, prosumers and consumers can trade electric energy directly without the need for central entities such as the utility companies (Shrestha et al. 2019). On the other hand, wholesale energy trading allows electric energy to act as a commodities tradable in the energy marketplace. For instance, prosumers can trade electric energy in the energy markets in the form of flexibility, carbon credit, guarantee of origin, among other energy assets (Green and Newman 2017).

1.2. Motivation and state-of-the-art

Generally, smart grids integrates various computing technologies, like IoT (e.g. sensors and actuators), distribution controls (e.g. boards and circuit breakers), and advanced metering infrastructure (e.g. smart meters). Such systems that integrates the physical and the cyber worlds are called Cyber-Physical Systems (CPS) (Dedeoglu et al. 2020). However, seamless integration between the physical (distribution controls) and the cyber (IoT) elements in DET faces various challenges such as privacy leakage, security vulnerabilities, and centralised control which leads to a single point of failure. For instance, in order to realise optimal demand response scheduling in residential, commercial and industrial buildings, smart meters are installed to collect near-real-time electricity consumption data (Guan et al. 2018). Nevertheless, the near-real-time consumption data may disclose the user's private information, which can be exploited by a cyber-attacker to pose an attack. Further, a cyber-attacker can compromise smart grids security by exploiting the IoT system vulnerabilities, which include software errors, protocol misconfiguration, and lack of security guards such as antivirus. The Saudi Aramco electric grid attack in 2017 (Alshathry 2016), and the Ukraine power grid attack in 2015 (Zetter 2016) are examples of smart grid cybe-attacks, where attacker exploited IoT devices' vulnerabilities in the smart grid system. Lastly, Shrestha et al. (2019) outlined the centralised control architectures that rely on large utilities for DERs monitoring and control as a single point of failure and the course of communication bottlenecks in DET systems (Shrestha et al. 2019).

Recently, blockchain technology has emerged as a viable solution to the aforementioned power grid challenges, due to its ability to provide decentralisation, immutability, security and privacy, and transparency in DET systems (Dorri, Kanhere et al. 2019). Blockchain derives its salient features from a suite of technologies, which include hash functions, consensus algorithms, cryptographic schemes, and smart contracts (Alharby and van Moorsel 2017). Although the initial application of blockchain was to transfer coins in the Bitcoin cryptocurrency network (Nakamoto 2008), its application extends beyond cryptocurrencies. For example, blockchain has been used in supply chains to provide trust and provenance; in healthcare to share medical records securely; and in property management to provide proof of ownership, among other use cases (Faculty, Engineering, and Others 2020). In the energy industry, blockchain is being used to provide the following DET solutions: (i) Transparency in wholesale energy trading system, such as billing and metering, automation, grid management, and sharing energy resources (Andoni et al. 2019). (ii) Collaborative local communities, including P2P energy trading, microgrids, and consumer-centred marketplaces (Dorri, Kanhere et al. 2019), and (iii) Provenance in e-mobility and vehicular energy networks (Andoni et al. 2019).

1.3. Our contributions

This work reviews the literature to identify barriers in blockchain-based DET systems. More specifically, our work presents the following set of contributions:

(1)	We propose an evaluation and review framework for analysing blockchain-based DET systems.
(2)	We present a systematic literature review methodology for conducting a search, screening, keywording, and mapping literature review in blockchain-based DET systems.
(3)	We present a state-of-the-art literature review of the blockchain-based DET studies conducted using the proposed DET framework and the review methodology. We present our review results organised into a taxonomy of open challenges, and proposed solutions for each barrier.
(4)	Lastly, we identify gaps and potential future research opportunities in blockchain-based DET system.

The rest of the paper is organised as follows: In Section 2, we present an overview of blockchain followed by the review of the related work. Section 3 draws on the systematic literature review methodology to analyse the relevant research on blockchain-based barriers in DET. Then, Section 4 expounds on the review findings identifying specific challenges under each barrier and the relevant blockchain-based solutions. Section 5 provides a discussion of identified gaps and open issues. Lastly, Section 6 culminates this Work.

2. Background and related work

This section gives a conceptual theory and basic principles of blockchain for applications in DET. First, we elaborate on the fundamental concepts of blockchain technology, including the basic models and algorithms. Next, we give an overview of the status quo in DET before integrating blockchain. Finally, we draw a comparative summary of related work on blockchain applications in DET to distinguish our review and the existing reviews.

2.1. Blockchain overview

Figure 2 shows the five main layers of a blockchain platform, namely data, network, consensus, contract, and application layers (Lu 2018). Each layer host a suite of technologies that collectively provide the blockchain's salient features. The data layer is the lowest and constitutes a distributed ledger, data structures, hash functions, and cryptographic algorithms. The network and consensus layers are the intermediary layers. The consensus layer mainly generates data blocks using a consensus algorithm, while the network layer broadcasts the blocks for verification by each node in the network. Finally, at the top are the contract and the application layers. The contract layer host the smart contracts execution engine and account models, such as unspent transaction outputs (UTXO), while the application layer includes the services and applications platforms. We review each layer in detail as follows.

Figure 2. Blockchain overview.

2.1.1. Data layer

A ledger is a decentralised database (or a shared state transactions log) shared among different participants' nodes (peers) transacting in a distributed P2P network (Andoni et al. 2019). Generated transactions in P2P networks are then bundled together into structures called blocks. Then, a hash function links adjacent blocks into a chain structure, as illustrated in Figure 3.

Figure 3. Blockchain data structure (chain of blocks).

Essentially, each block has a body and a header. A block header has multiple information fields, including a hash of the previous block, a root hash of the Merkle tree, and a timestamp indicating when the block was generated. The Merkle root is generated by iteratively hashing pairs of transactions stored in the block's body (Burkhardt, Werling, and Lasi 2018). Each transaction contains several fields, including date, time, and information about the transaction participants. A blockchain network participant can securely and efficiently use the Merkle root to verify the authenticity and non-repudiation of the transactions.

2.1.2. Network layer

This layer contains two main components: an unstructured P2P network and a broadcast protocol. An unstructured P2P network interconnects nodes in a blockchain network such that nodes have the freedom to join or leave the network. Since unstructured P2P networks have no fixed topology, each node has a routeing algorithm to verify and transmit blocks through the broadcast protocol (Lv et al. 2002). A broadcast process allows a block to spread throughout the network in seconds, reducing the chances for collusion. The transacting node propagates the generated transactions block to multiple neighbouring nodes in the blockchain network. On successful block validation, the neighbouring nodes continue the propagation process and return a success message to the transacting node, which verifies the information passed. If the validation process fails, the neighbouring node discards the propagation process and returns a rejection message to the transacting node (Lu 2018).

2.1.3. Consensus layer

The consensus layer deals with enforcing network rules that describe how nodes should reach a verdict on the transactions' validity in a propagated block (Lu 2018). The consensus process does not require the sender to trust any network participants. Greater decentralisation increases trust and transparency because of more complex consensus mechanisms. Conversely, more centralised decision-making led to quicker consensus, at the cost of reduced transparency and increased centralisation of trust. Figure 4 illustrates the degree of decentralisation for various network models (Burkhardt, Werling, and Lasi 2018).

Figure 4. Network models.

In a centralised network model, only one node controls the ledger. The node is assumed to be highly trusted. The decentralised network model adopts a client-server architecture, which partitions the processing workload (e.g. execution, validation, storage, etc.) among the resource providers, called servers, and service requesters, called clients. Next, the P2P network model partitions the workloads between nodes such that nodes in the network are equally privileged (i.e. peers). Blockchains adopt the P2P network model. However, in some applications, we may require data confidentiality. Thus, a P2P network is not feasible. Consequently, some blockchains, such as Hyperledger Fabric, have introduced channels to cluster the P2P network. Although all peers are equally privileged, the channels act as confidentiality boundaries.

Based on the trust models mentioned above, blockchains are currently classified into three categories, detailed as follows:

(i)	Public blockchains (permission-less): Public blockchains are blockchain platforms where any node in the network can play an integral role in the consensus process. The prevalent consensus algorithm in permission-less blockchains is the Proof-of-work (PoW), whose initial introduction was in Bitcoin (Nakamoto 2008). Although the PoW consensus mechanism has a high degree of decentralisation, it imposes high computation and communication costs for generating and propagating blocks, a process called mining (Li et al. 2019).
(ii)	Private blockchains (permissioned): permissioned blockchains are blockchain platforms where only nodes from one organisation can determine the final consensus. The Practical Byzantine Fault Tolerance(PBFT) (Castro 2001) and Raft (Howard and Crowcroft n.d.) are some examples of the permissioned consensus algorithms used in private blockchains. Private blockchains have lower computation and communication costs than public blockchains but at a reduced degree of decentralisation.
(iii)	Consortium blockchains (partially permissioned): Several organisations construct a consortium blockchain platform, so only nodes from the selected organisations would be part of the consensus process (Zheng et al. 2017). The consortium increases the degree of decentralisation compared to the private blockchains.

2.1.4. Contract layer

The contact layer is composed of a rules engine that executes the terms of the agreement stipulated in a smart contract. By definition, a smart contract is a self-executing piece of code that starts a contract's execution between individuals. Smart contracts need to run in a deterministic manner to achieve consensus. Blockchain systems that guarantee transaction correctness, authenticity and compliance through the execution of smart contracts include, the Hyperledger Fabric (Androulaki, Barger et al. 2018), Ethereum (Wood 2018), Tendermint (Kwon 2014), and Quorum (Sankar, Sindhu, and Sethumadhavan 2017)). Therefore, the contract layer offers various advantages such as reducing fraudulent transactions, enabling trust-less arbitration, implementing enforcement costs, and minimising exceptions − both malicious and accidental (Bashir 2017).

2.1.5. Application layer

The application layer hosts the end-user's batch and online applications that utilises the blockchain services. A batch processing application manages massive data amounts executed routinely. In contrast, an online processing application collects activities and issues simultaneously (Jena n.d.). Example of online applications include web, mobile, and API gateways.

Having discussed some of the main technologies underpin blockchain, we introduce the DET concepts and highlight some challenges that need blockchain-based solutions in the following section.

2.2. DET overview

The power grid in many countries is undergoing disruptive changes due to the emergence of DERs and increasing consumer-led electricity demand. ARENA (ARENA 2019) identified various scenarios of the possible consumer-led response to decrease electricity demand and increase DERs uptake, as illustrated in Figure 5. The identified scenarios are outlined as follows:

• Set and forget: In this scenario, the future DET models will recognise the busy lives of many buildings occupants. Therefore, managing peak demand in smart buildings and smart homes will be on a 'set and forget basis'.

• Rise of prosumer: In this scenario, prosumers ranging from residential, industrial, to commercial will actively engage in the energy market but with a stronger preference for collective generation, consumption and storage. Therefore, future markets such as the P2P energy trading market must provide transparency in scheduling and optimisation techniques for a trusted consumer-centric DET model.

• Leaving the grid: In this scenario, prosumers are expected to go off-grid. Prosumers have different storage locations, like small-scale batteries at the prosumer premise and mobile storage on EVs. They can also engage in feed-in tariffs (through reverse metering enable by smart meters) by selling their excess generation back to the grid.

• Renewables thrive: energy storage plays a larger part in the entire electricity system. Therefore, energy transactions and usage data should be available for better decision-making. But ensuring that privacy is preserved on the shared data remains a key research challenge.

Figure 5. Possible pathways for electricity system transitions in Australia.

Introducing more DERs in the power grid at the distribution level presents numerous challenges to the electricity system. (Karimi et al. 2016) reviewed various studies and documented the impact of high penetration of DERs in the electricity grid. For example, the increase in residential air conditioning and the growth in solar PV generation during the day have resulted in a typical duck curve profile in multiple global areas. As seen in Figure 6, high generation from solar PV results in a reduction in the grid consumption during the day; however, the evening peak demand due to the use of loads such as air conditioning is not reduced, resulting in a duck-shaped consumption profile.

Figure 6. Energy consumption in megawatts at a different time of the day illustrates ‘The Duck curve problem’. Source: (Green, Martin, and Cojocar 2018).

Renewable Energy Sources (RES) have widely been adopted as alternative energy sources as they provide economic and environmental benefits for sustainability development (Dang et al. 2019). However, energy generation from RES is stochastic and distributed, making it hard to manage energy demand-supply balancing effectively. Additionally, DET undermines the economic fundamentals of traditional grid management and regulations predicated upon a centralised grid utility (Green, Martin, and Cojocar 2018). We review each of these challenges below.

2.2.1. Flexibility trading

Energy production from RES, such as wind and solar, depends on the weather, making it massively volatile and unpredictable (Mohandes et al. 2019). Besides short-term changes, RES shows mid-term (daily, weekly) and long-term (seasonal) changes. Hussain, Mullapathi Farooq, and Selim Ustun (2019) illustrated the requirement for adequate resource utilisation characterised by demand flexibility activities, like quick response and inadequate production, to manage RES volatility. In general, the literature is rich on Demand Side Management (DSM) prerequisite to continuing RES penetration. DSM is the ability to change the patterns and strength of end-user electric energy demand through incentives, and energy efficiency guidance measures, among other energy management activities (Hussain, Mullapathi Farooq, and Selim Ustun 2019). Some everyday energy management activities, including energy efficiency, spinning reserve, and demand response, as demonstrated in Figure 7. These activities influence demand and energy response (Palensky and Dietrich 2011).

Figure 7. DSM activities.

In a case study, Khatoon et al. (2019) examined blockchain technology more closely to make energy efficiency markets more transparent and secure. Pop et al. (2018) suggested a blockchain-based decentralised initiative for handling demand response concepts in a tamper-proof method. Finally, a spinning reserve is the market model correlated to the available energy in virtual storage, usually meant to recoup for power failure or occurrence fall (Ferreira and Lucia Martins 2018). Madhu, Vyjayanthi, and Modi (2019), proposed using electric vehicles managed through Blockchain to issue ancillary services like a spinning reserve. These activities utilise load balancing, scheduling, and other DET strategies. In DET markets, players exchange assets (e.g. money, electricity, green certificates, power purchase agreements, etc.) and value (e.g. demand response, sustainability, efficiency, cost-saving, etc.) on digital marketplace platforms. However, trust plays a critical role in interactions between distributed entities in such networks. Traditionally, trust is achieved through trusted third parties such as banks, retailers, or government organisations (Yu et al. 2018). Therefore, market growth in transactive networks (or TEMs) requires transitioning trust management from centralised to distributed control.

2.2.2. DET opportunities

High penetration of DER in the network, such as storage, electric vehicles, and onsite generation, has opened new ventures to provide current services and business setups in the electricity market. These technologies have placed prosumers at the centre of transition. For example, governments worldwide provide subsidies to incentivise solar PV owners to export their energy to the grid through a Feed-in tariff. Prosumers have greater control and autonomy of energy use and can earn revenue for dispatching their electricity to the grid. The P2P energy trading business concept is an attractive option for deriving more value from these distributed energy sources. The emergence of DER has also opened up opportunities for aggregators, who can package large bundles of individual DER sources and participate in the wholesale market. Aggregators can deliver grid flexibility through P2P models. For example, aggregators could control a cluster of air conditioners to manage network peak demand. Various technical studies have established the viability of trading DER for grid management. (Wang et al. 2015) reviewed electricity markets in three regions for DER and DR participation. Eid et al. (2016) reviewed market design options for DER trading in the electricity markets.

Although DET promises to be a potential disruptor, questions remain about reliability, security and transparency of the available energy technologies. As outlined in Section 1.2, blockchain technology is emerging as a novel technology that promises to address trust, security, privacy, and transparency in the emerging energy technologies in the smart grid ecosystem (Dorri, Kanhere et al. 2019). Next, we review how blockchain is being used to address these challenges.

2.3. Related work

Various articles have reviewed blockchain-based applications in DET (Wang, Zhou et al. 2019; Abdella and Shuaib 2018; Andoni et al. 2019). Wang, Zhou et al. (2019), reviewed blockchain incorporation in various DET topics. First, they studied existing solutions based on a distributed architecture. Then, they investigated security and privacy solution. Finally, they reviewed energy trading schemes classified into three categories: transaction object matching, energy cost minimisation, and auction pricing mechanism. The authors concluded that some challenges, such as lacking a regulatory system and policies, still need to be solved. Abdella and Shuaib (2018), reviewed existing research on blockchain applications in P2P DET. The review identified blockchain scalability, setup cost, computational efficiency, transaction latency, security and privacy, among others, as key technological challenges in realising blockchain-based P2P DET. Lastly, Andoni et al. (2019) reviewed 140 blockchain-based research projects and distributed applications (DApps). In their conclusions, the authors argued that (i) blockchain technology needs to offer scalability, speed, and security before its mainstream adoption in the energy industry. (ii) although blockchain systems may realise significant cost savings by eliminating trusted authorities, they might not have a competitive advantage as they currently have high deployment costs. (iii) both regulatory and legal spheres act as administrative barriers to adopting blockchain technology. (iv) the lack of blockchain systems interoperability and standardisation also significantly slows the adoption of blockchain in DET.

Meanwhile, after surveying the literature related to blockchain application in DET, we identified a comprehensive list open challenges using the taxonomy of barriers introduced by Zabaleta et al. (2020). We classified these open challenges under technical, standardisation, administrative, trust, and economic barriers, as summarised in Table 1. The taxonomy grouped the barriers as follows:

• Scalability, privacy security and interoperability challenges as technical barriers.

• Reliance on trusted third parties as a standardisation barrier.

• Lack of regulatory policies as administrative barrier.

• Grid management and optimisation challenges as economic barriers.

Table 1. A summary of the reviewed blockchain-based DET reviews.

		Reviewed barriers
Reviews	DET topics covered	Technical	Standardisation	Administrative	Trust	Economic
Wang, Zhou et al. (2019)	• Energy Transaction	✓	✗	✗	✗	✓
	• Consensus Mechanism
	• System Optimization
Abdella and Shuaib (2018)	• Demand Response	✓	✗	✗	✓	✓
	• Power Routing
Andoni et al. (2019)	• Grid Management	✓	✓	✗	✗	✓
	• Carbon Trading
	• Electric E-mobility
Zabaleta et al. (2020)	• Local Flexibility Markets	✓	✓	✗	✓	✓
	• Local Energy Markets
Villar, Bessa, and Matos (2018)	• Distributed generation	✗	✗	✗	✓	✓
	• Self-Consumption
	• Demand Response
	• Electric Vehicles
Our work	• Market Segments	✓	✓	✓	✓	✓
	• Value Propositions
	• Value Chains
	• Market Mechanisms

While the reviews above have outlined the benefits and barriers to applying blockchain in smart grids, none has gone beyond conceptualising the potential solutions to the outlined open challenges. This paper further investigates potential solutions to the identified challenges in the smart grid design space, specifically at the intersection of DET and CPS. In the following section, we give a preliminary of systematic review methodology and the proposed framework used for the review analysis.

3. Preliminaries

Preliminary to our work, we first analyse the systematic mapping methodology used to review the blockchain-based DET studies in this area. Then, we present a detailed description of the proposed review framework.

3.1. Methodology

Since DET is still an emerging area with no well-defined structure, we adopt the systematic mapping study method presented in Petersen, Vakkalanka, and Kuzniarz (2015). The systematic mapping study is a light version of the well-known systematic literature review methodology (Kitchenham et al. 2009). As illustrated in Figure 8, it has five steps, described below:

Figure 8. The systematic mapping process.

3.1.1. Research questions

Our Work explores the Blockchain incorporation in DET to show barriers hindering its widespread adoption in smart grids. Therefore, in this Work, we seek to solve the following research questions

RQ1:	What are the technical complexities affecting blockchain adoption in developing market segments for DET?
RQ2:	What are the standardisation and administrative barriers hindering blockchain-based value proposition in the energy markets?
RQ3:	How can we integrate grid utilities into the new value chains without incurring high trust costs?
RQ4:	What are the economic barriers hindering the optimal value creation in DET markets?

3.1.2. Search strategy

We followed the keyword search strategy suggested by Kitchenham and Charters (Kitchenham and Charters 2007) in their PICO (Population, Intervention, Comparison, and Outcomes) guideline. After identifying the keywords, we decided on the scientific databases to conduct our search. Table 2 shows the search string and the results from the databases of IEEEXplore, ACM Digital Library, Science Direct, and Scopus. Our focus was only to include peer-reviewed papers published in conferences, journals, workshops, symposiums and books published in the last five years.

Table 2. Minimal requirements for each blockchain role.

Database	Search string	Results
IEEE	(‘All Metadata’:blockchain) AND (‘All Metadata’:energy OR ‘All Metadata’:power OR ‘All Metadata’:electricity) AND (‘All Metadata’:trading OR ‘All Metadata’:market) AND (‘All Metadata’:building OR ‘All Metadata’:‘Smart home’)	19
ACM	[All: blockchain] AND [[All: ‘energy’] OR [All: ‘power’] OR [All: ‘electricity’]] AND [[All: ‘trading’] OR [All: ‘market’]] AND [[All: ‘buildings’] OR [All: ‘smart home’] OR [All: ‘residential’]] AND NOT [All: ‘build’] AND [Publication Date: (01/01/2017 TO 12/31/2021)]	28
ScienceDirect	(‘blockchain’ AND (‘power’ OR ‘energy’ OR ‘electricity’) AND (‘trading’ OR ‘market’) AND (‘building’ OR ‘smart home’) AND prosumers)	1716
Scopus	( TITLE-ABS-KEY ( blockchain ) AND TITLE-ABS-KEY ( energy OR power OR electricity ) AND TITLE-ABS-KEY ( trading OR market ) AND TITLE-ABS-KEY ( building OR ‘smart home’ OR residential))	89

3.1.3. Screening for relevant papers

Paper screening excluded studies that were not relevant to answering the research questions. First, if the search returned many studies and many were identified as outside the scope, we added more keywords to the search string, such as ‘prosumers’, which is relevant to the topic. Next, we excluded studies such as magazines, PowerPoint presentations, grey literature, posters and lecture notes that did not report empirical findings or literature. Then, applying inclusion and exclusion criteria based on the title and abstract proposed by Dybå and Dingsøyr (2008) was conducted to exclude papers irrelevant to our study. For example, the search query returned papers on building as a verb rather than a noun. Figure 9 shows the number of included articles at each stage.

Figure 9. Screening process.

3.1.4. Keywording using abstract

Keywording ensures that we only consider the existing studies relevant to our research question (Marew, Kim, and Hwan Bae 2007). In this study, we follow the keywording technique described by Alharby and van Moorsel (2017). As illustrated in Figure 10, we first read through the abstracts looking for keywords and concepts that reflect any relevant component in our conceptual framework. Then, we used those keywords to classify the papers into various categories.

Figure 10. Classification scheme.

3.1.5. Data extraction and mapping

Data extraction and mapping is the last step in the systematic mapping study process, used to map all the required articles to the specific research questions in this study. We developed complete data extraction and synthesis template, as given in Table 3, to document the data extraction process in an Excel table. For each article, we first identified the study parameters. Second, we noted the challenges focussed on by the authors. Next, we summarised the proposed solution. Then, we concluded with the paper's findings.

Table 3. Data extraction and synthesis template.

#	Scheme	Description
1	No.	This is a record serial number
2	Reference	This field contains the article's title, authors, and other metadata
3	Study parameter	Study parameters are used to describe the entire population being studied. For example, we want to know the target customers.
4	Focus	Seeks to answer the question; which specific challenges are the author aiming to address?
5	Proposed solution	Refers to the proposed methods and tools for solving the identified challenges/gaps.
6	Findings	A comparative summary of the research findings.

3.2. DET conceptual framework

In economics, the relationship structures among participants determine the classifications of markets. DET is a new market structure that delivers value-added services within the energy distribution network. Based on Mengelkamp et al. (2018) and Block, Neumann, and Weinhardt (2008), we derive four market components that represent an abstract (generic) characterisation of a DET energy market. Figure 11 gives a schematic overview of the main components that make up the DET conceptual framework. We describe each component as follows.

Figure 11. Components of a DET conceptual framework.

3.2.1. Market segments

A market segment is the group of customers to whom traders ultimately target their products or services (Orlov 2017). Borrowing from the global social networks and service-oriented systems in Camarinha-Matos (2016), Camarinha-Matos introduced the term 'Collaborative Networks (CNs)' to represent market segments where both the customers and traders join competencies in sharing resources. In DET, research studies identify various CNs, their motivation and the technology and mechanisms to form their network. Examples include Virtual Power Plants (VPP) (Seven et al. 2020), local collaborative communities (Ahmadi et al. 2015), Microgrids (Goranovic et al. 2017), and Vehicular Energy Networks (VEN) (Goranovic et al. 2017). From the technical side, the electric grid is a network of sub-networks. The broad establishment of C.N.s issues a foundation for establishing Local Energy Markets (LEM) and Local Flexible Markets (LFM). DERs can participate in LEM and LFM coordinated through IoT, artificial intelligence, and blockchain technologies. LEM enables electricity trading between participants in low voltage markets, whereas LFM enables trading in medium to high voltage digital marketplaces, mainly serving more extensive distribution network needs (Zabaleta et al. 2020). For instance, LEMs include microgrids and residential collaborative communities, while VPPs and VENs are examples of LFMs.

Although market segments differ in size from the traditional electricity market, their design reflects the centralised market structure, relying on a TTP (Zabaleta et al. 2020). While blockchain promises to address the issues of TTP, the proposed decentralised approaches suffer from a lack of scalability and often lead to confidential user information leakage. These and other technical barriers limit the widespread adoption of blockchain in smart grids. In Section 4.1, we review proposed solutions for these and other technical barriers that limit the widespread adoption of blockchain in smart grids.

3.2.2. Value proposition

The value proposition defines how value is created, communicated and delivered to the customer (Richter 2012), often through products and services. In the energy context, the value propositions of a classical utility comprise the generation and supply of electricity for a fixed cost per kilowatt-hour and emissions credits equal to the issued cap limit level. However, the value of clean energy technologies extends far beyond the ability to generate clean energy or avert GHG emissions. Pater (2006), outlined additional categories of value propositions to consider when valuing clean energy technologies. Rather than compartmentalising the values, the categories provide ‘buckets’ such as risk management, social benefits, policy incentives, emission reduction, and flexibility, among others, into which we can group similar values for a given scenario.

Blockchain can capture value propositions through tokenisations, transforming the value into crypto-assets (Baum 2020). However, crypto-assets face administrative and standardisation barriers due to the lack of regulation, legislation and policies governing how they transact within the energy markets. In Sections 4.2 and 4.1.4 we review the administrative and technical barriers hindering the adoption of blockchain in DET and their proposed solutions.

3.2.3. Value chain

A value chain describes a set of activities performed by an establishment or establishments in a specific industry to deliver the value proposition to customers (Orlov 2017). In the energy industry, the global electricity value chain creates value (delivers electricity) through a set of activities that begins with generation, moves through transmission and distribution, and ends with consumption. The arrow bars in Figure 12(a) show the activities carried out in a centralised energy market value chain. However, with the incorporation of Information Communication and Telecommunication (ICT) in the classic grids, the global value chain has been disrupted as energy and information can flow in both directions. Consequently, new DET value chains have emerged, as illustrated by the horizontal bars in Figure 12(b).

Figure 12. Energy trading value chains. (a) Traditional Centralised Energy Trading system. (b) Decentralised Energy Trading systems.

However, in DET, trust is costly to maintain. Trust is a fundamental precondition underpinning the exchange and economic coordination of a market (Davidson, Novak, and Potts 2018). In simple terms, a cost of trust is referred to as ‘trust cost’. Therefore, as ICT has extended the sphere of consumers in the global energy value chain with the ability to generate and trade energy over the Internet, trustworthiness has come under stress among participants in the DET value chains (Shyamasundar and Patil 2018). Although the invention of blockchain seems to have addressed the trust issue with distributed control, transparency, and provenance, the available blockchain frameworks incur high communication and computation cost. The trade-off manifests through trusted third parties in private and consortium blockchain for conflict resolution. In Section 4.3 we review the trust barriers hindering the adoption of blockchain in DET.

3.2.4. Market mechanisms

A market mechanism refers to the trends that industries and other market entities use to create value for their participants. Market mechanisms ask the question: When is the optimal value created during trades? In energy trading economics, examples of market mechanisms include game theory, incentives, and off-sets. Sorin, Bobo, and Pinson (2019) and Parag and Sovacool (2016), identified two umbrella market models for blockchain-based DET: peer-to-peer and pool-based market. Additionally, in our previous work (Karumba et al. 2020), we proposed a new collaborative market model. Table 4 gives a short description of these market models as follows:

Table 4. Summary of the three market models.

Market model	Description	Revenue optimisation	Relationship mode
	Facilitate direct transactions between buyers and sellers without relying on a TTP	Competition among(1) Individual agents	Prosumers compete amongst themselves
	Facilitates blockchain trading to serve the interests of many (for example, a group of prosumers can form a community or organisation)	Collaboration within(1) Prosumer group(2) Distribution grid	Prosumers act as a utility to provide smart grid services such as flexibility to the grid.
	Serve the interest of prosumer groups and utility services such as demand-supply balancing.	Strategic co-opetition among(1) Prosumer group(2) Local/main grid(3) Service providers	Prosumer groups compete with utilities over clients but collaborate in grid services

(a) P2P market model: A P2P market model is characterised by simultaneous multi-bilateral trades among participants along a predefined trading scheme (Sorin, Bobo, and Pinson 2019). In principle, the predefined trading scheme should facilitate direct transactions between buyers and sellers without relying on a TTP, as illustrated in Section 3.2.4. However, using the term ‘peer’ does not necessarily mean that the buyers and seller are in the same physical location but rather directly connected in the market communication graph.

(b) Pool-based market model: Participants in pool-based market design tend to pool their resources together for mutual benefits (e.g. cost-sharing, market sourcing, etc.), which individuals could not achieve otherwise. As illustrated in Table 4, the pool-based network topology is more organised than the P2P network topology. Outside the energy sector, pool-based market structures are more mainstream, with Uber¹ as a common example of a sharing market structure (Belk 2014). Within the energy sector, a pool-based market model would be suitable for community-organised prosumer groups such as island microgrids and virtual power plants (Dang et al. 2019). Theoretically, prosumers could pool their prosumption resources to generate revenue. Alternatively, prosumers could use small and medium-scale companies that act as aggregators or energy services providers to pool their energy resources. Given the centralised nature of decision-making, a centralised DLT system such as a permissioned blockchain, where only nodes from one organisation determine the final consensus, could be used to implement its value chain. Such systems require market mechanisms such as fair scheduling for optimal value creation.

(c) Collaborative market model: In our collaborative market (Karumba et al. 2020), we considered that different market mechanisms competing for common resources might come together to achieve a common goal collaboratively. Regardless of the group, each node acts as an autonomous market participant. For example, a rooftop solar panel system and battery storage in the same building owned by one individual may act independently in the DET market but contribute to demand-supply balancing in the overall grid. Therefore, we need to dynamically coordinate the complex interactions between the network nodes as a value proposition. First, the collaborative market value chain considers all interaction factors such as ownership, location, time, and context, which form clusters of closely related nodes referred to as collaborative communities. Second, the collaborative network structure offloads computationally intensive tasks to trusted nodes in the same cluster. Lastly, given the heterogeneity of smart devices, incentive mechanisms are adopted to incentivise device participation for optimal value creation.

In market mechanisms, however, blockchain frameworks face economic barriers such as high sustainability costs (i.e. computation and communication costs), demand-supply balance, and grid defection. In Section 4.4 we review these economic barriers and their proposed solutions.

4. Blockchain barriers in DET: a systematic review

In this section, we used the conceptual framework presented in Section 2 to analyse the blockchain-based DET articles mapped in Section 3.1. After reviewing the knowledge, we identified a comprehensive list of challenges hindering blockchain adoption in transactive energy networks (digital marketplaces). We classify these challenges within the following main themes: technical, administrative, standardisation, social, and economic barriers. The screened articles have been reviewed in this section and summarised in Tables 5 and 6.

Table 5. Technical barriers to blockchain-based DET and the currently proposed solutions.

Market segment
Technical barriers	Proposed solutions
(a) Scalability	• ‘A Lightweight Scalable Blockchain for IoT security and anonymity’ (Dorri, Kanhere et al. 2019);
	• ‘Blockchain-Based Scalable and Tamper-Evident Solution for Registering Energy Data’ (Pop et al. 2019)
(b) Privacy	• ‘Distributed meter data aggregation framework based on Blockchain and homomorphic encryption’ (Wang, Luo et al. 2019);
	• ‘A Blockchain-Based Privacy-Preserving Payment Mechanism for Vehicle-to-Grid Networks’ (Gao et al. 2018)
(d) Security	• ‘Secure Vehicular Network using Advanced Blockchain’ (Hassija et al. 2020);
	• ‘A Security Model for Electric Vehicle and Charging Pile Management Based on Blockchain Ecosystem’ (Huang et al. 2018);
	• ‘A blockchain-based secure mutual authentication with fine-grained access control system for industry 4.0’ (Lin et al. 2018);
	• ‘Blockchain-based charging coordination mechanism for smart grid energy storage units’ (Baza et al. 2019);
	• ‘A blockchain-based decentralised model for IIoT’ (Kumari et al. 2020)
(e) Interoperability	• ‘Blockchain interoperability framework with trusted data transfer’ (Abebe et al. 2019).
	• ‘Channels: Horizontal scaling and confidentiality on permissioned blockchains’. (Androulaki, Cachin et al. 2018)

Table 6. A summary of administrative, standardisation and economic barriers to blockchain-based DET and the currently proposed solutions.

Value proposition
Administrative barriers	Proposed solutions
(a) Regulations	• Regulations are still required in designing microgrid energy markets (Mengelkamp et al. 2018) and P2P DET (Abdella and Shuaib 2018)
	• ‘Planning and Governance for Decentralised Energy Assets in Medium-Density Housing: The WGV Gen Y Case Study’ (Green, Martin, and Cojocar 2018).
Value chain
Standardisation barriers	Proposed solutions
(a) Trusted third parties	• ‘Distributed ledger technology for document and workflow management in trade and logistics’ (Wang et al. 2018).
	• ‘Blockchain-enabled workflow operating system for logistics resources sharing in E-commerce logistics real estate service’ (Li, Shen, and Huang 2019).
	• ‘BPMN modeling and reference guide: understanding and using BPMN’ (White and Miers 2008)
	• ‘Distributed ledger technology for document and workflow management in trade and logistics’ (Wang et al. 2018)
	• ‘Blockchain-enabled workflow operating system for logistics resources sharing in E-commerce logistics real estate service’ (Li, Shen, and Huang 2019) ‘Model-Driven Development Patterns for Mobile Services in Cloud of Things’ (Cai et al. 2018)
	• ‘On Using Blockchain Based Workflows’ (Bore et al. 2019)
Market mechanism
Economic barriers	Proposed solutions
(a) Operating cost	• ‘Implementation of Blockchain technology for Energy Trading with Smart Meters’ (Hussain, Mullapathi Farooq, and Selim Ustun 2019)
	• ‘Demand Side Management: Demand Response, Intelligent Energy Systems, and Smart Loads’ (Palensky and Dietrich 2011)
	• ‘Blockchain in energy efficiency: Potential applications and benefits’ (Khatoon et al. 2019)
	• ‘A Novel Framework for Monitoring Solar PV based Electric Vehicle Community Charging Station and Grid Frequency Regulation using Blockchain’
	• ‘A Novel Debt-Credit Mechanism for Internet of Vehicles’ (Liu et al. 2019)
	• ‘Prosumer integration in wholesale electricity markets: Synergies of peer-to-peer trade and residential storage’ (Zepter et al. 2019).

4.1. Technical barriers

Based on the screened articles, the technical barriers stemming from current blockchain technology limitations include scalability, security, privacy and interoperability (Kouhizadeh, Saberi, and Sarkis 2021). The technical analysis was conducted below and summarised in Table 5.

4.1.1. Scalability

Open issues: Most IoT devices, in the intelligent grids, use resource-constrained sensors, controllers, and actuators (Abdella and Shuaib 2018). Therefore, it can be difficult for such devices to participate in blockchain networks for application in DET. Additionally, whereas current work focuses on the feasibility of developing lightweight blockchains, the conventional blockchains follow the order-execute architecture, which consumes significant resources of the participating devices (Androulaki, Barger et al. 2018). In the order-execute architecture, special nodes called miners first bundles transactions into an ordered block through a resource-intensive process called mining. Then the mined block is broadcast to all nodes where all activities face execution and validation through the consensus protocol such as PoW. Consequently, blockchains based on order-execute architecture are (i) computationally intensive, (ii) have high bandwidth consumption and processing time, and (iii) consume significant storage size over time (Karumba et al. 2020).

Proposed solutions: (Dorri, Kanhere et al. 2019), issued a Lightweight Scalable Blockchain (LSB) for applications in IoT environments such as the brilliant home setting. LSB incorporates a lightweight consensus mechanism, which minimises the execution overhead of authenticating new blocks. The framework focuses on the blockchain nodes building trust through direct or indirect evidence. Similarly, Pop et al. (2019) presented a scalable blockchain-distributed ledger that combined distributed queuing systems with a NoSQL database. Their process minimises the validation frequency of energy activities on the Blockchain when engaging with real-time energy information gathered by smart meters.

4.1.2. Privacy

Open issues: The decentralisation feature requires all the blockchain network nodes to preserve a duplicate of the immutable ledger (Dorri, Hill et al. 2019). However, replicating personal data such as payment records, fine-grained energy generation, and consumption data may leak sensitive information. For example, payment records can reveal user identity and location, while fine-grained energy data can reveal usage patterns raising privacy concerns. Privacy is the constraint on data or information adopted to ensure that personal and sensitive data are not disclosed to unauthorised parties (Gai et al. 2019). Although blockchain uses pseudo-identities to preserve user privacy, it is still susceptible to linking attacks. In linking attacks, attackers obtain identifiable information from the distributed ledger by linking it with other external datasets (Gai et al. 2019).

Proposed solutions: To address the privacy challenges, various solutions (Wang, Luo et al. 2019; Gao et al. 2018; Dorri, et al. 2021) have been proposed. Wang, Luo et al. (2019) leveraged the use of homomorphic encryption (HE) to enhance blockchain security in their meter data aggregation framework. The proposed framework aims to combine power utilisation data of end-users such as commercial or industrial buildings from various geographical locations. In this case, the authors used HE technology to encrypt the smart meter data to protect individual plain-text meter data privacy during the aggregation process. Gao et al. (2018) presented a blockchain-based privacy-preserving payment mechanism based on the Elliptic Curve Digital Signatures algorithm (ECDSA) to address the disparity between data privacy and data sharing in V2G networks. By leveraging the ECDSA in their registration process, the new payment mechanism enabled value-add service providers to analyse payment records without inferring private information about E.V.s, such as their identities. In a recently study, Dorri, et al. (2021) proposed the use of a removable ledger to minimise the volume of activities stored on the Blockchain. Their architecture introduced a Temporary Chain that stores transactions only for a particular period to enhance the Privacy of users and data while increasing blockchain scalability, throughput, and low latency.

4.1.3. Security

Open issues: The most common security threats in DET systems are confidentiality, integrity, and availability (CIA) (Dedeoglu et al. 2020). Although blockchain could address the CIA security threats, its adoption for application in IoT is limited by its resource-constrained nature (Karumba et al. 2020; Dorri, Kanhere et al. 2019). IoT vendors and users are often concerned with the functionality and pricing of IoT devices while ignoring their security configuration (Lin et al. 2018). Figure 13 shows a taxonomy of CIA security threats of blockchain-based DET systems.

Figure 13. BDET system attack taxonomy.

In a confidentiality threat, the attacker gain access to unauthorised confidential data either during collection, transmission or storage in distributed ledgers (Hassija et al. 2020). For instance, an attacker may have forged identities to impersonate authorised devices (e.g. sensors, edge nodes, cloud nodes or a user) in blockchain-based DET systems to gain access to confidential data. These attacks that threaten data confidentiality by compromising identities are referred to as authentication attacks, including replay, Sybil, and impersonation attacks, as illustrated in Figure 13.

When data and control signals are sent and received by IoT devices, edge, cloud, or user terminals, no falsification should occur during the transmission or storage. Integrity threats imply that an unauthorised party accesses and modifies data or controls information. Under the integrity threats category can find False Data Injection (FDI), poisoning, forgery, and data tampering attacks (Lee et al. 2020). FDI attacks are common integrity attacks where an adversary modifies the stored measured states or directly manipulates the measured states in DET systems (Qi et al. 2017). In forgery attacks, an adversary forges control information to deceive the end-users or actuators in a DET system to act accordingly (Lee et al. 2020). In replacement attacks, an adversary violates data integrity by replacing original data or control signatures with corrupted data or signal signature (Lee et al. 2020).

Finally, in availability attacks, an attacker's goal is to make services offered by individual components of a DET system, including IoT devices, edge and cloud servers, unavailable (Lv et al. 2019). First, due to the high volume and veracity of the IoT data, a centralised cloud system is used to store the actual data. Data centralisation, however, exposes the DET systems to the Distributed Denial of Service (DDoS) attacks orchestrated by the Mirai Botnets on centralised cloud servers (Cui and Guin 2019). Second, a large-scale compromise of IoT devices facilitated by a Botnet can elucidate availability threats. Other common attacks launched by botnets include FDI attacks, and Electrical Fault Attacks (EFA) (Lv et al. 2019).

Proposed solutions: (Huang et al. 2018) investigated an authentication mechanism to counter confidentiality attacks in V2G networks. The mechanism leverages the proposed lightning network and smart contracts model for secure registration, scheduling and authentication phases of EVs and charging stations. Similarly, Lin et al. (2018) proposed a novel blockchain-based system for secure mutual authentication of Industry 4.0 applications dubbed BSeIn. BSeIn integrated the salient features of blockchain with: (i) attribute-based signatures to autonomously authenticate terminals; (ii) message authentication codes to efficiently authenticate gateways; and (iii) certificate-less multi-receiver encryption for secure mutual authentication.

To address the availability threats brought about by the centralised nature of cloud services, various authors (Baza et al. 2019; Kumari et al. 2020; Hill et al. 2021) have proposed blockchain-based approaches for services coordination in a decentralised manner. Baza et al. (2019) proposed a blockchain-based framework for coordinating Energy Storage Units (ESU). Blockchain was used to ensure the availability of the requested data by resisting tampering and DDoS attacks in a single charging coordinator. Motivated by the decentralised nature of blockchain technology, Kumari et al. (2020) proposed a blockchain-based decentralised model for IoT data dissemination in industrial IoT dubbed DMIIoT. Through a case study, the authors then evaluated the efficacy of DMIIoT based on data load balance, energy management cost, and transmission delay parameters. Finally, Hill et al. (2021) proposed the BlockTorrent protocol to share private data between multiple stakeholders while addressing data availability and accessibility challenges.

Further, Fan, Liu, and Zeng (2020) utilised the underpinning characteristics of blockchain to propose a Decentralised Privacy-preserving Data Aggregation (DPPDA) scheme for smart grids. DPPDA uses a Boneh-Lynn-Shacham short signature and SHA-256 function to ensure data integrity. Another study by Li et al. (2020) proposed a secure transmission and storage framework that can resist theft and forgery attacks. The framework implements five algorithms for intelligent data sensing, dividing sensed data into blocks, encrypting and transmitting, signature, and verifying stored data. Similarly, Mbarek et al. (2020) addressed the issue of data tampering in distributed transaction transmission. The authors proposed an enhanced blockchain mechanism using mobile software agents that control and detect seller nodes' malicious activities to counteract possible FDI attacks.

4.1.4. Interoperability

Open challenges: Today, there is no ‘de facto’ standard for interoperability between DET applications (DApps) in the wide spectrum of blockchain protocols (Abebe et al. 2019). The existing blockchain protocols have different information and consensus handling methods. Therefore, DApps do not provide cross-value transfer even when implemented on the same blockchain protocol, leading to data and information silos (Abebe et al. 2019). Thus, interoperability becomes the new barrier in the foundation of blockchain-based DET value chains.

Proposed solutions: (Robinson 2021) surveyed cross-chain communications models, providing them based on the top-level usage scenarios they attempt to meet, including value swapping, cross-chain messaging, and blockchain pinning. Androulaki, Cachin et al. (2018) used channels to address the interoperability challenge. Their interoperability architecture used a trusted channel as a trust intermediary to transfer value between two channels. Abebe et al. (2019) presented an interoperability architecture that uses relays for communications between permissioned blockchain networks. Dorri, Luo et al. (2019) introduced atomic meta-transactions in their proposed secure private Blockchain to extinguish the TTP incorporation in cross-chain trades.

4.2. Administrative barriers

Blockchains suffer from administrative uncertainties in various countries. For instance, regulations in many countries do not recognise blockchain-based DET markets such as P2P energy trading. Hence, the governance of blockchain-based DET markets requires modification of market standardisation and regulation policies.

4.2.1. Regulation policies

Although blockchain could help overcome many energy-related issues, there are no clear regulatory frameworks (Programme, Postgraduate Studies n.d.). More specifically, it is unclear how to employ taxation on risk management, social benefits, emission reductions, and all possible industry-wide value propositions traded as crypto-assets. Regulation policies are rules to achieve certain aims or goals. Policymakers in ‘crypto-friendly’ jurisdictions enact policy changes to attract blockchain investment or technology-enhanced entrepreneurship in blockchain-based DET (Novak 2019). Some of these crypto-friendly jurisdictions include Australia, Estonia, some states in USA, Singapore, and Switzerland. Conversely, policymakers in ‘crypto-unfriendly’ countries have responded to the blockchain-based DET by prohibiting or significantly raising the cost of its applications (Mengelkamp et al. 2018).

4.3. Standardisation barriers

Open issues: This theme focuses on establishing trust among participants in DET market segments. Balta-Ozkan et al. (2013) investigated social barriers that prevent the adoption of new smart grid technologies such as smart homes. After an in-depth review, they indicated inadequate trust in intelligent technologies as the primary barrier. Participants inquired if standardisation services and technologies would provide sustainable development in the future because of the profit incentive by technology generators and energy suppliers. The participants also suspected that the utility companies (or other TTP) would not pass on the consumer's benefits. For instance, consider the Australian energy market scenario depicted in Figure 14. A consumer-retailer price dispute is not resolved by the participants but by the Energy Consumer Australia² (ECA) and Energy Retailers Association of Australia (ERAA) umbrella organisations, respectively. ECA and ERAA negotiate to resolve a dispute between themselves, or if negotiations break down, the Australian Energy Market Commission³ (AEMC) or a court of law may intervene. Given the conflicting interest, ECA and ERAA mutually distrust each other. The two participants follow an agreed-upon process when negotiating a dispute. The process defines simple terms such as who proposed the unit price, who performed what and when etc. However, depending on the trust cost (payoffs), they may not have equal incentives to follow this process. If the national energy retail price decreased, ERAA might be less forthcoming in responding to negotiation meeting dates. Conversely, if AEMC increased the retail prices, ECA may intentionally stall the process to continue enjoying the low prices.

Figure 14. High-level consumer-retailer workflow for an Australian energy market.

Proposed solutions: The terms of collaboration between distrusting participants is an example of a business contract, while executing the agreement between participants is the workflow (California, Southern 2018). Therefore, the system that supports the execution of a distributed workflow (workflow in a distributed system) must provide two guarantees: (i) Workflow correctness − meaning that the workflow must enforce the agreed-upon contract, such that no party obtains an advantage by acting out of turn or fails to fulfil an obligation; (ii) Immutable consensus history − the system should provide transaction provenance or an audit trail that can decide which party did violate the agreed-upon contract when required. Wang et al. (2018) argued that blockchain could guarantee the correctness and suitability of distributed contract execution. Within their blockchain framework, they defined contracts and workflows as follows:

(a) Contract model: According to Li, Shen, and Huang (2019), a contract describes the whole business logic that enables various participants to work on tasks to achieve business goals. For example, the trade contract between a consumer and a retailer depicted in Figure 14 may be defined by a set of rules outlined in Table 7.

Table 7. Consumer-retailer trade contract.

Steps	Business process
1.	Consumer C request X kilowatts hours of energy from Retailer R.
2.	Retailer R commit to supply X kilowatts hours of energy to Consumer C.
3.	Consumer C to pay amount Y when X kilowatts hours of energy is received from Retailer R.

They integrated the business process statements with the workflow concept to achieve better logical expression depending on the modelling purpose (e.g. execution or representation).

(b) Workflow model: There is a need for a management mechanism to enforce accountability in the business process (Zou 2016). Following the consumer-retailer example, you may recall that in case negotiations break down, AEMC is used as a trusted third party to provide interventions by executing the dispute resolution workflow illustrated in Figure 14. Therefore a workflow model was used as a contract management mechanism. However, to eradicate the trusted third parties' incorporation, the articles in Li, Shen, and Huang (2019) and Wang et al. (2018) proposed a blockchain-based workflow management concept that guarantees correctness and auditability of distributed contract execution. Li, Shen, and Huang (2019) suggested the use of Petri-net (Zou 2016) methodology for representation workflow modelling and event-condition-action rules (Cai et al. 2018) methodology for executable logical expressions. Preferably, Wang et al. (2018) opted for the event-condition-action rules (Cai et al. 2018) methodology, which can model workflow representation and executable logical expression using notations such as JSON or XML. Additionally, Bore et al. (2019) outlined three key concepts that developers must present in a workflow: Actor (participants), documents (i.e. logical expressions), and the Action list (execution sequence). This study uses Petri-net-based workflow abstraction models for high-level DET workflow models. Figure 14 illustrate a Petri-net-based workflow model for enforcing the consumer-retailer contract without needing interventions from AEMC (trusted third party).

4.4. Economic barriers

This category generally involves the study of market obstacles to the given value proposition. Drawing particularly on Good, Ellis, and Mancarella (2017) we describe the various market barriers, including hidden costs, pricing or system value, and margins. Chai and Yeo (2012) gave an extensive list of market barriers. The hidden costs include initial and maintenance investments associated with participants in the market. Then the system value is associated with costs related to the new contracts carried by the new markets.

Open issues: With the most commonly used PoW consensus method, miners are incentivised using several built-in incentive methods to verify transactions on the blockchain. For example, miners who verify transactions by solving a puzzle receive a bitcoins reward in the Bitcoin blockchain. This reward is periodically cut in half, and once bitcoins mints all coins, miners will rely on transaction fees as the only source of revenue. Thus, public blockchains are the least preferred blockchain in DET, as they are counter-intuitive to the sustainable energy development goal described in the Kyoto protocol (Loeser and Treede 2008). In blockchain-based DET markets, the mining process also increases the energy cost to cover transaction fees.

Proposed solutions: (Wang, Zhou et al. 2019) provided a general evaluation of blockchain-based pricing schemes to address the transaction fee challenges. In their work, they argued that the having more adaptable pricing updates in DET applications, the system can issue the best matching concept, which is very important in market democratisation to encourage renewable generation investments. Similarly, several studies (Liu et al. 2019; Zepter et al. 2019; Hahn et al. 2017; Thakur and Breslin 2018; Sabounchi and Wei 2018; Tripathi et al. 2016) have examined market mechanisms proposed to address pricing and buyer-seller matching issues in blockchain-based energy trading. Liu et al. (2019) proposed a motivation-based investing and pricing mechanism in their EVs data trading system to encourage loaning among vehicles. Their mechanism formulates a two-stage Stackelberg game to maximise the borrower and lender's profits jointly. We could adopt this pricing mechanism for other energy attributes such as electricity, carbon credits, and DERs. Similarly, Noor et al. (2018) proposed a game-theoretic approach incorporating storage components to capture every individual customer's needs. Additionally, their system suggested smoothing the dips in the load profiles ignited by supply constraints. Zepter et al. (2019) introduced the Smart elecTricity Exchange Platform (STEP) for interfacing between the prosumer societies and wholesale electricity markets. STEP used a two-stage stochastic model to authenticate the suggested pricing concept and analyse its financial potential.

5. Open research challenges

The blockchain-based energy trading research solutions reviewed in this article show that there has been a steep rise in the adoption of Blockchain for various services and applications in energy trading. There are dozens of blockchain-based DET trial projects in production environments and hundreds in early development phases focussed on wholesale markets, P2P energy trading, EVs charging, and energy attribute certificate (EAC) trading. These trial projects have iteratively found problems with adopting blockchain in DET markets and made corresponding improvements. However, several issues need to be resolved to realise the long-term value of blockchain in DET.

5.1. Data management issues

The distributed ledger is one of the critical characteristics of Blockchain, where a massive amount of information is replicated on each network node to allow transparency and visibility in DET (Dorri, Hill et al. 2019). However, IoT devices do not have enough computational and storage capacity, which derails blockchain incorporation on a production scale (Panarello et al. 2018). Therefore, blockchain data management techniques should be investigated further, including storage optimisation and increasing the available storage capacity through distributed cloud storage.

5.2. Blockchain interoperability issues

As highlighted above, hundreds of blockchain-based energy trading frameworks are designed for transactive energy space. Several blockchain platforms on which developers implement these frameworks exist, ranging from popular ones such as Ethereum, Hyperledger Fabric, and IOTA to completely developing new ones such as Lition (2019) and Enosi (n.d.). However, implementing these frameworks on different blockchain platforms complicates how they will interoperate and share value among participants transacting across different multi-chains. In our previous work (Karumba et al. 2020), we investigated this issue and proposed a hypergraph-based adaptive consortium blockchain model for privacy-preserving cross-chain transaction verification. Inter-chain operations are still nascent and require efforts to prevent double-spending attacks in critical electronic payments for energy trading (Aitzhan and Svetinovic 2018). Additionally, double spending transactions in the case of unbundled renewable energy certificates (i.e. energy and renewable certificates sold separately) still occur where renewable generators sell the same REC multiple times. Therefore, standards and mechanisms for interoperable blockchain architectures are required to combat these challenges and expand inter-market opportunities.

5.3. Economic incentivisation issues

The current blockchain incentive model is unsustainable for high transaction volumes in DET systems. Block creation in current blockchain systems involves a costly mining process, which increases trust costs. Cumulatively, these costs sometimes raise the renewable energy prices in collective self-consumption communities to more than the retail market prices making it hard to justify how the use of this technology outweighs the price benefits. In light of this, project developers need to design more value capture market mechanisms in addition to financial incentives tailored around the benefits of data integrity, enhanced security and actionable intelligence.

5.4. IoT security issues

Although blockchain has numerous applications in the energy sector, its suite of technologies incurs high computation and communication costs, hindering its application in resource-constrained environments (Jindal et al. 2020). Smart grids are inherently resource constrained IoT networks, which limit the application of blockchain in the energy sector (Karumba et al. 2020). Consequently, researchers in the IoT domain have proposed the use of: removable ledgers (Dorri, et al. 2021), scalable networks (Karumba et al. 2020), and lightweight consensus mechanisms (Dorri, Kanhere et al. 2019; Biswas et al. 2020) for effective blockchain application in IoT networks. While the outlined research works have significantly reduced the computation and communication costs in blockchain, the efficiency of its cryptographic scheme has not been fully explored for applications in IoT.

5.5. Data privacy issues

In a book chapter published in the Wireless Blockchain: Principles, Technologies and Applications (Imran et al. n.d.), we highlighted the privacy paradox in blockchain. One of the blockchain's salient feature is visibility and transparency (Dorri, Hill et al. 2019), which is achieved by giving each party an equal opportunity in transaction data verification and storage. However, transaction data such as payment and energy usage records may contain personal information, like identities, locations, and energy production and consumption patterns, raising privacy and security concerns (Karumba et al. 2020). Privacy-preserving techniques currently being used include Zero-Knowledge Proofs, Ring Confidential Transactions, Elliptic Curve Digital Signature Algorithm (ECDSA), and mixing techniques, all of which can be incorporated to enable transactions data privacy. However, these techniques increase communication and computational complexity and consequently increases the blockchain's energy footprint. Therefore, research topics such as secure off-ledger (Dorri, Hill et al. 2019) data storage techniques should be explored further to figure out how transactions data privacy in blockchain could be preserved by off-ledger or on-ledger storage without losing the salient features of blockchain.

6. Conclusions

This paper evaluates the technical, administrative, standardisation, and economic barriers hindering the application of blockchain in DET. Commencing with an examination of the definitions and concepts of electricity markets, we encapsulate the pivotal components, participants, and technologies of DET markets. After that, we presented a comprehensive market analysis framework for conducting a systematic literature review on blockchain-based DET. The reviewed literature showed that, although Blockchain provides its salient features of decentralised trust, immutable provenance, visibility, and transparency in DET, its data management, standardisation, and incentivisation mechanisms still require improvement for mainstream adoption. Simultaneously, there is a need to increase scalability while reducing computation and communication costs without security and privacy trade-offs. Our review, therefore, gives the DET market stakeholders a generic characterisation to examine the relationships of the different market components, namely market segment, value proposition, value chain and market mechanisms. In summary, the relevance of this study is to guide organisations, managers, policymakers, and other stakeholders toward the mainstream adoption of blockchain in the DET ecosystem. With insights from our research, the stakeholders can prioritise their efforts in resolving particular difficulties hindering Blockchain's integration in DET.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 https://www.uber.com/.

2 https://energyconsumersaustralia.com.au/.

3 https://www.aemc.gov.au/regulation/regulation-1.