Business process modeling helps organizations to capture their workflows visually as diagrams that could be then used to share best practices, identify inefficiencies in ongoing activities, instruct employees, use them as reference solutions, etc. Design and analysis of business process models are essential technologies of the Business Process Management approach, which is successfully adopted and practiced nowadays by many large and medium enterprises. Therefore, business process models should be considered as organizational assets that depict usable and competitive business solutions, which value could be proven by benchmarking. Single reference business process models or even collections of business process models are already accessible on Internet, sometimes for free or, usually, for purchasing because of the value of transferred knowledge. However, peer-to-peer exchange of business process models on a commercial basis is still far from a unification. At the same time, the tremendous growth of the cryptocurrency market and the adoption of Bitcoin by governments (first by El Salvador in June 2021) makes crypto-economics, also referred to as “tokenomics”, usable for organizational knowledge sharing and exchanging without third party authorities, such as banks, or payment systems. Moreover, collaborating parties could reach a consensus when exchanging business process models using smart contracts and crypto-tokens to access shared knowledge artifacts. Therefore, this paper proposes an approach to business process model tokenization using blockchain technology and smart contracts. There was proposed as an Ethereum smart contract that combines features of the business process model collection and the non-fungible token. A prototype of a decentralized application was developed and its usage was demonstrated.
Nowadays digital transformation is a trend in enterprise management. In the first place, digital
transformation is associated with Business Process Management (BPM) and its applications in
business process modeling, and automation using BPM suites. However, business process modeling is
used not only to draw executable workflows – they are usually simplified enough and describe mostly
routine document flows. The main goal of business process modeling includes a visual representation
of business activities as graphical diagrams to identify and understand ongoing workflows, find
bottlenecks for improvement, and ensure communication between IT (Information Technology) staff
and business stakeholders. Widely used reference models of typical enterprise business processes are
used by organizations to adopt and tune concerning industry and internal needs. At the highest levels
of BPM, maturity organizations have in their possession large collections of business process models
that are extremely valuable for them and their competitors. Therefore, successful and time-proven
business scenarios captured as process models could be sold to other organizations willing to achieve
positive BPM outcomes. Secure and peer-to-peer exchange of business process models could be
organized using blockchain technologies, including cryptocurrencies and smart contracts. The only
problem of business process models tokenization, i.e. representation of them as digital tradable assets
on a certain crypto-platform [
This paper is structured as follows: in the introductory section is given an overview of related work (sub-section 1.1) and the problem statement is made (sub-section 1.2); section 2 outlines the approach to business process models tokenization based on smart contracts usage; section 3 contains results and their discussion regarding decentralized application prototype development and validation; conclusion and further research plans are outlined in section 4.
In general business processes are considered as structured collections of manual or automatic (by
IT systems, e.g. BPMS, CRM, ERP, and others) executed activities necessary to achieve business
goals and satisfy end customers [
An extensive classification of business process modeling perspectives was given by J. Krogstie in
[
Even though EPC notation is still in use in the area of business process modeling, many users have
replaced it with BPMN since it is a standard [
Pools and lanes are used to model process resources: pools define business parties, i.e. boundaries
of a business process, while lanes define roles, i.e. organizational units or persons, that take part in a
business process execution [
Sample BPMN model of a goods purchase business process, which outlines considered symbols, is demonstrated in Fig. 1 below.
This diagram contains one pool “Goods purchase” that defined business process boundaries, three lanes “Supply department”, “Warehouse”, and “Accounting department” that define responsible roles for the process execution. As for core BPMN symbols, there are: start event “Out of stock” that starts process instance, intermediate events (e.g. “Goods received” and “Order placed”), end event “Goods supplied” that completes process instances, tasks (e.g. “Place order”, “Receive goods”, “Store goods”, and “Verify invoice”), split and join AND-gateways to implement parallel execution of “Store goods” and “Verify invoice” tasks (see Fig. 1).
The term “blockchain” means an immutable or read-only data structure – a linked list of blocks, a
directed acyclic graph (DAG), or a tree-like data structure, in which new data can be only appended at
the end of the blockchain [
There are the following principles of blockchain technology that make it adaptable in many industries where security, transparency, and consistency are necessary features: decentralization, peerto-peer communication, transparency, pseudonymity, irreversibility, and computational logic [11]. In general, blockchain is useful only when several entities are collaborating and sharing data since local copies are maintained on participating nodes to ensure consistency and tamper resistance [11].
The underlying idea of blockchain is using cryptography to link data blocks of referred distributed and decentralized chains. The first block is called the “genesis block”, it does not refer to any previous block in the chain [12]. However, each block contains data (usually as strings), nonce (the unique number usually related to mining – none generated repeatedly until meets consensus requirements of new block adding), and the hash value of all fields of the previous block (i.e. the cryptographic link to a preceding block, also referred as “previous hash”), the hash value of all fields of the current block [12]. The original consensus algorithm Proof-of-Work (PoW) was used to confirm transactions and add new blocks to the Bitcoin blockchain. This algorithm assumes miners (persons who share their computational power to support the network) are competing to confirm transactions and get rewarded in cryptocurrency [11]. Today PoW is used by such most popular cryptocurrencies like Bitcoin and Ethereum. However, Ethereum is planned to be switched to another popular consensus algorithm referred to as Proof-of-Stake (PoS). This algorithm overcomes high energy use since transactions are confirmed by validators who stake coins instead of sharing their computing resources. The higher stake is, the higher the probability to add a new block with pending transactions and getting rewarded [11].
However, in comparison to Bitcoin and other cryptocurrency-oriented blockchain platforms (such as Dogecoin, Bitcoin Cash, Litecoin, etc.), Ethereum was designed as a “world computer” with the ability to host smart contracts – programs to be executed within a blockchain platform [13].
Bitcoin as the first cryptocurrency was proposed in 2008 by Satoshi Nakamoto, who is presumably an anonymous person or group of persons hiding by this pseudonym. The difference from traditional payment systems is that electronic currency could be transferred in a peer-to-peer manner without a central party that needs to check the records of ownership. Blockchain technology helps to prevent the double-spending problem and enforce the validity of transaction records [13]. One more crypto-asset “token” is almost the same as the cryptocurrency – a bearer instrument used to transfer value between parties over the blockchain network; whereas tokens are created by a single party to represent a certain value, cryptocurrency is generated by the network as the reward for miners or validators [13].
Token technology was introduced and standardized by Ethereum and its smart contracts, which code describes how each token should work [13]. However, Ethereum was planned as a globally distributed computing network that uses publically stored immutable programs – smart contracts also referred to as decentralized applications [14]. However, decentralized applications usually include a client-side created using markup, style sheets, and JavaScript (JS). Decentralized applications (or DApps) use the Web3 JS library to interact with the smart contracts deployed on the blockchain [12]. DApps are immutable (as any data stored on the blockchain) and perform exactly as they were developed, without the possibility of fraud, downtime, censorship, or interference [14].
In the context of blockchain, tokenization means a representation of real physical or electronic assets digitally on the blockchain [15]. There could be commodities, real estate, ownership rights for arts or other collectibles, currency, or any other kinds of assets. As advantages tokenization offers faster and cheaper transaction processing, flexibility, decentralization, security, and transparency [15], but there are disadvantages, such as regulatory and legality issues, as well as technical barriers caused by the use of DApps [15]. There are two types of crypto-tokens [16]: Fungible Tokens (FT), which value is identical among all tokens, and which are exchangeable to each other (i.e. one FT token could be replaced by another FT token similarly to digital cash). Non-fungible Tokens (NFT) are unique and not equal to each other in value (i.e. each NFT token is different from others and cannot be replaced by any of them).
Another explanation of FT and NFT given in [16] says that dollar bills are exchangeable to other dollar bills, so they are fungible, whereas baseball cards or other collectibles are unequal in their value and cannot be replaced with another one.
From the technological point of view, both FT and NFT are implemented as smart contracts, which are programmed in a specific way. Several smart contract standards for tokens were developed by the Ethereum community [17]. The most popular contract standards are [17]: ERC20 standard for fungible tokens, which is suitable for multiple use cases, such as payment tokens, loyalty coins, gift cards, etc. A great example of NT implemented as the ERC20 token on the Ethereum blockchain is Tether USD, a stable cryptocurrency (also referred to as the “stablecoin”) that digitally represents USD (United States Dollar) [18]. ERC721 standard for non-fungible tokens, which is suitable for documents, land titles, digital identities, real estate, collectibles, etc. A great example of NFT implemented as the ERC721 token on the Ethereum blockchain is CryptoKitties, an Ethereum-based game in which players buy, sell, and breed collectible digital cats [19].
Nowadays OpenZeppelin organization provides extensive references for the development of ERC20, ERC721, and other smart contract standards [20]. 1.2.
The problem of blockchain-driven cross-organizational business process modeling, versioning, and executing was considered by Härer in [21] and Fill in [22], Hull in [23], Viriyasitavat and Hoonsopon in [24], De Sousa and Corentin in [25], Milani and Garcia-Banuelos in [26], and others. Thanks to the considered blockchain advantages, inter-organizational storage of tokenized business process models provides collaborative parties with proof of authorship, censorship resistance, timestamping, and immutability. Smart contracts and token standards provide decentralized financial (DeFi) capabilities, such as peer-to-peer exchange of enterprise knowledge presented as business process models without the need of any third-party authorities, i.e. banks or payment systems. Therefore, the problem of business process model tokenization remains relevant and respective information technologies should consider the latest trends in blockchain technology and digital economics. Since in Spring 2021 NFTs got a focus in global crypto attention, mostly for collectibles and digital art (the entire market exceeds 130 USD by Spring 2021) [27], it seems like a great opportunity to extend the use cases of NFTs with tokenization of enterprise models, in particular – business process models.
The NFT better suites process models that are unique and unequal in terms of their syntactic and
semantic properties, which could be used to define the value of shared models. There could be used
SEQUAL (Semiotic Quality) framework [
As for the blockchain platform, there could be chosen Ethereum as the pioneering and still leading smart contracts platform despite its competitors, such as Binance Smart Chain, Polkadot, Solana, and others [28]. Thus, the ERC721 standard may be used for business process model tokenization. As it is shown in Fig. 2, the URI (Uniform Resource Identifier) of a business process model should be used as the token URI. Besides the token URI, the ERC721 standard should be extended with syntactic and semantic features of business process models. After NFT is minted (i.e. published on the blockchain), syntactic and semantic features are used to evaluate the validity and completeness of a process model to define its value for collaborating parties (Fig. 2).
According to [
In terms of formalisms proposed in [
Formally, syntactic quality is defined as following [
# M \ L M missing , # M where: # M \ L is the number of business process model M statements that do not correspond to the BPMN language L ; M missing is the number of missing statements that make a model syntactically incomplete (i.e. missing sequence flows or events); # M is the total number of business process model statements.
As for semantic properties, there are also validity and completeness phenomena [
In [
# M \ D Semantic completeness 1
, , # M # D \ M # D (1) (2) (3) where: # M \ D is the number of business process model M statements that do not belong to the modeling domain D ; # D \ M is the number of model M statements that are not correct or relevant to the modeling domain D ; # D is the number of statements in the modeling domain.
Therefore, a quality-oriented smart contract that may be used to assess tokenized business process
models should consider formalisms of SEQUAL framework (1), (2), and (3) regarding syntactic and
semantic quality of business process models given in [
Nisnyvnatlid , Nisnycnotmplete ,
N isnevmalid , N isnecmomplete , where:
Nisnyvnatlid is the number of syntactically invalid statements in a business process model; Nisnycnotmplete is the number of syntactically incomplete statements in a business process model; Nisnevmalid is the number of semantically invalid statements in a business process model; Nisnecmomplete is the number of semantically incomplete statements in a business process model.
In addition to (4) and (5), there should be noted the total number of statements N , which could be considered as equal to # M and # D , since the syntax and semantics of real business process models are completing each other in order to reflect business activities.
Finally, there are following mapping should be noted:
Syntactic : tokenId Nisnyvnatlid , Nisnycnotmplete , Nisnyvnatlid Nisnycnotmplete N , (6) Semantic : tokenId
Nisnevmalid , Nisnecmomplete , Ninvalid Nisnecmomplete N ,
sem
Total : tokenId N , (8) where tokenId is the unique identifier of each tokenized business process model.
Whereas (6), (7), and (8) define the structure of business process model properties, there are following metrics should be calculated then:
Syntactic validity 1 N isnyvnatlid ,
N Nisnycnotmplete ,
N N isnecmomplete . (4) (5) (7) (9) (10) (11) (12)
Semantic validity 1 N isnevmalid , N
N
Interpretation of calculated metrics (9) – (12) may be done using the Harrington scale [29] to transform crisp values of syntactic and semantic validity and completeness into linguistic values. To visualize obtained linguistic values, it is proposed to use “Green”, “Yellow”, and “Red” color codes for “Good”, “Satisfied”, and “Bad” values respectively (see Table 1).
Table 1
Good Syntactic validity 0.8,Syntactic completeness 0.8, Green
Obtained evaluation results will help collaborating parties define values and formulate prices of shared business process models for further exchange.
Syntactic validity 0.63, Syntactic completeness 0.63,
Syntactic validity 0.63, Syntactic completeness 0.63,
As it was outlined before, the ERC721 NFT standard [30] will be used for business process model tokenization. According to the problem statement, the ERC721 standard should be extended with the following behavior: model publishing and contacts data publishing (so parties can reach each other for collaboration). Before a model is published to the blockchain, i.e. the respective NFT is minted, special metadata JSON (JavaScript Object Notation) file should be published to the Internet and be accessible by a certain URI (later will be used as the token URI). Such JSON document accessible by the token URI might include the following properties [30]: Name (i.e. a brief description of a depicted business process). Description (i.e. a more detailed description of a depicted business process). Image URI (i.e. an image of a BPMN diagram).
Obviously token metadata stored in JSON files will not be stored on the blockchain. However, such an approach is considered the most efficient from the perspective of transactions’ speed and cost [31]. Whereas to save tamper resistance provided by the blockchain technology, only the hash value of an entire document could be found using the secure algorithm (e.g. SHA-256 or others) and stored on the blockchain to verify the identity of the original document [31].
Then, to summarize, we may define two more mappings in addition to formalisms (6), (7), and (8) denoted earlier:
Party : address contacts, contacts UTF8 , ,
Hash : tokenId sha256tokenURI , (13) (14) where: address is the user address in the Ethereum network; contacts is the user’s contact information outside the blockchain (e.g. email address, phone or messenger number, etc.) represented as the string value of UTF-8 characters set UTF8 or even the empty characters set (i.e. the empty string) if a party decided to keep pseudonymity; sha256 is the hashing algorithm SHA-256 applied to the JSON document tokenURI .
Moreover, besides the (6) – (8) and (13) – (14) mappings, the smart contract should store a number of models modelsCount used both as the capacity and the identifier of next minted token. In general, the minting or model publishing sequence diagram may look as following (see Fig.3).
Thus, as depicted in Fig. 3 behavior should be implemented in the form of an ERC721-compatible smart contract that allows for the publication and read tokenized business process models and related data. In terms of UML class diagrams, the smart contract can be depicted as follows (see Fig. 4).
Contract “ERC721” with the dashed border line represents OpenZeppelin ERC721 implementation provided at [30] (see Fig. 4), which is used as the generic for developed contract “EtherBPMNNFT”. The smart contract also contains two structures “SyntacticStatements” and “SemanticStatements” that serve as tuples of semantic and syntactic characteristics of tokenized BPMN models. As shown in Fig. 4 mappings of “EtherBPMNNFT” smart contract implement formalisms (6) – (8) and (13) – (14), while functions “publishModel” and “publishContacts” are used to mint NFT tokens and store collaborator contact information respectively. Some of ERC721 standard functions can be called from the developed contract, these functions are also mentioned in Fig. 4.
Considered smart contract is barely useless without having a special decentralized application that allows ordinary users to interact with the blockchain. The following use cases should be supported by the DApp prototype for BPMN model tokenization (see Fig. 5).
The system architecture of the decentralized application is similar to any client-server web application, whereas instead of an application server and persistent storage the smart contract and blockchain ledger is used respectively [12]. The structure of the DApp prototype is shown in Fig. 6.
Components with dashed border lines are third-party libraries, such as Bootstrap for user interface, CryptoJS for SHA-256 calculation, Web3 for interaction with the Ethereum platform, Ethereum blockchain ledger itself, MetaMask Web3 provider (as the Google Chrome extension), and also some Ethereum mobile, desktop, or web wallet that can be used to hold business process models in the form of ERC721-compatible NFTs, which are tradable and exchangeable as any other digital crypto-assets. 3.2.
Developed prototype of a decentralized application for business process model tokenization that can be used only with MetaMask or another Web3 provider. In the authors’ opinion, MetaMask is the easiest and the most popular tool to work with Ethereum networks, either the main network or test network, since it only requires the installation of the Google Chrome extension. According to the sequence diagram (Fig. 3), the tokenization procedure starts with the storing of the JSON metadata and BPMN diagram image in the third-party non-blockchain storage. Let us imagine we are intending to store BPMN shown in Fig. 1. It is image is already given and could be stored somewhere on the Internet with the permanent URI. As for other metadata properties, the following could be given: Name: “Goods purchase”. Description: “The BPMN model of a goods purchase business process”.
Therefore, respective JSON (as well as the BPMN diagram image) could be stored in the GitHub repository of this project and may look like the following (see Fig. 7).
The URI of future NFT is following – “https://raw.githubusercontent.com/andriikopp/blockchainrepository/main/nft/models/0.json”. Now, after the non-blockchain data is stored, let us fill and submit the model publishing form in the DApp (see Fig. 8).
When the “Publish” button is pressed, the MetaMask appears with the request to sign a transaction as it is shown in Fig. 9. In case of the transaction is successfully mined, a respective token will be minted and can be accessed by the address “0xabc33640b17def441cfb455efa1c3f8f490f4616” and ID “0” (since it is the first NFT in a collection). A respective token could be added to MetaMask for future exchange with other parties (see Fig. 9).
Now BPMN diagram shown in Fig. 1 can be kept in the MetaMask wallet or any other Ethereum wallet that supports ERC721 tokens, exchanged with other Ethereum users, or even traded using NFT exchanges and marketplaces like other crypto-tokens that represent collectibles or digital art.
The developed DApp prototype also displays tokenized business process models with all given JSON-based data and owner’s address, but also with the specific characteristics of BPMN diagrams: quality metrics of syntactic and semantic validity and completeness, a hash value used to ensure the identity of business process model metadata, and owner’s contact information (if it was preliminary published to the smart contract, of course) as it is shown in Fig. 10.
According to Fig. 10, each tokenized BPMN model are displayed syntactic and semantic metrics that correspond to (9) – (12) formulas and color codes given in Table 1. Besides metrics, it is possible to reach the owner by retrieving its contact information (in case it was provided) and checking identity by comparing the SHA-256 hash value (stored on the blockchain in the smart contract mapping) to the actual SHA-256 hash value of JSON requested by token URI.
The form of contacts information publishing, as well as examples of requested contact information and SHA-256 hash of the NFT data, are shown in Fig. 11.
Provided quality metrics may help to define the value of the business process model when exchanged, while the hash value may help to check whether the BPMN has been tampered with by the previous owner. Currently, the DApp is at the software prototype stage being under construction and testing [32]. The MetaMask or another wallet with an in-built Web3 provider should be installed to use the DApp, as well as the Ropsten testnet account should be present for basic usage.
In this research paper, we proposed the approach to business process model tokenization based on blockchain technology and smart contracts. Based on the performed state-of-the-art overview, the BPMN business process modeling notation was chosen to describe tokenized business process models as the most widely used and considered standard in the BPM industry. Essentials of blockchain technology were considered to prove the relevance of business process model tokenization, and the essentials of smart contracts and decentralized applications were overviewed as well. Two standards of tokens – ERC20 and ERC721, which represent fungible and non-fungible tokens respectively were considered to select the appropriate token standard for BPMN diagrams. Based on features of NFTs, the ERC721 standard has been chosen, as well as Ethereum as the pioneering and still leading smart contracting platform has been chosen for implementation.
To provide tokenized business process models with specific features useful to define their value
for exchanging and trading, essential syntactic and semantic quality metrics were used: validity and
completeness [
Developed DApp prototype allows to publish a model as the NFT, review already published NFTs, including model names, descriptions, BPMN diagram images (preview and full size), quality metrics, owner addresses, request owner contact information (if provided), check the identity of tokenized BPMN models, and share own contact information if necessary. The smart contract has been deployed to the Ropsten Ethereum test network, while the DApp is under development and testing.
Future work in this area includes the development of the decentralized marketplace and exchange for BPMN models as NFTs, as well as a more rigorous evaluation of tokenized business process models using special methods and algorithms, rather than human judgment. The DApp should evolve into the full-scale ecosystem of tokenized BPMN diagrams that could be traded and swapped in the same way it could be done with cryptocurrency and NFTs nowadays.