The design of blockchain-based applications requires today in-depth technical knowledge of the underlying technologies and software frameworks. In order to investigate how enterprise modeling approaches can aid in designing such applications and aligning their structure and behavior with business needs, we conduct a comparison of two types of blockchain platforms using the ArchiMate modeling language. Based on a use case for Non-fungible Tokens for digital image licensing, we derive models for a software application using public and permissioned blockchain platforms. This permits us to gain rst insights into the adequacy of ArchiMate for representing blockchain-based applications and for highlighting the architectural di erences of public and permissioned blockchain approaches from a conceptual modeling perspective.
Prior work shows the use of enterprise modeling approaches such as ArchiMate
in di erent contexts related to blockchains. For example, ArchiMate has been
used previously in combination with the Business Model Canvas approach for
modeling blockchain-based business models and according enterprise
architectures [
As a foundation for the modeling of applications on blockchain platforms, we
designed a business use case employing Non-fungible Tokens (NFT) for managing
copyrights and licenses for digital images. An NFT is a representation of a digital
asset, attesting the uniqueness of the asset. Thus, NFTs of digital assets cannot
be interchanged. In the Ethereum ecosystem, ERC721 is the currently accepted
standard for implementing NFTs [
The use case considers two parties: a photographer and a licensee. The
photographer may create an NFT for an original digital image, identifying the holder
as the copyright owner. This 'copyright token' allows the holder to emit an
arbitrary number of licenses per image. Other parties may buy such a license for an
adjustable unit price, thereby obtaining the right to use the asset. A license is
represented as a fungible ERC20 token [
Traditionally, a photographer would submit the photo to an agency, managing and selling licenses on their behalf. Introducing a blockchain-based solution yields multiple bene ts: { No need for a central authority: Licensees buy directly from the copyright holder, eliminating costs associated with an intermediary party. { Traceable licenses: It is apparent who has bought a license token at what point in time and thus who holds a license. { Guaranteed uniqueness of the digital asset: Copyright tokens in the form of NFTs are unique and thus certify the uniqueness of the digital asset. No two parties can hold an NFT for an asset.
For verifying that an application for this use case is technically feasible using the ERC721 and ERC20 token standards, we implemented two prototypes for Ethereum and Hyperledger Fabric1. 4
From the prototype we derived the necessary application structure and infrastructure components. We then approached the creation of the models and the comparative study using an exploratory research approach. To this end, we found ArchiMate to be a promising candidate among the popular enterprise modeling languages. Thus, the NFT use case was represented using the ArchiMate business layer, while the application and technology layers are based on insights gained from the prototype development. Upon several iterative revisions and discussions between the authors, the model shown in Figure 1 nally emerged. We chose two popular blockchain platforms, the permissionless Ethereum and the permissioned Hyperleder Fabric platform, as represented in the ArchiMate technology layer. This choice was motivated by the signi cant di erences between these platforms as we will brie y outline in the following. 4.1
Ethereum is a popular programmable blockchain platform powering the cryptocurrency Ether. Ethereum's ability to execute smart contracts, i.e. pieces of code, enables versatile business applications and use cases. The public Ethereum network is permissionless. Anyone can participate by running a node contributing computing resources to the peer-to-peer network. Such participation is awarded by block rewards and charging fees on smart contract execution and transactions. 4.2
Hyperleder Fabric was developed for business blockchain applications. Ethereum and Hyperledger Fabric share fundamental blockchain concepts, such as executable code in the form of smart contracts - known as chaincode in Hyperledger Fabric, decentralized networking, transaction ledgers and a consensus mechanism. Contrary to Ethereum, the network is permissioned, i.e., all participants are authorized and know of each other. Participating roles are assigned roles by an authority, based on a network de nition speci c to the industry use case. 1 The prototypes are available via Zenodo: https://doi.org/10.5281/zenodo.
5211569
Business Layer Application Layer
The ArchiMate language provides a multitude of options for modeling a layered
enterprise architecture. By aligning business and IT layers, an understanding
of the dependencies between hardware, software and business services is
established. In principle, a blockchain application can be integrated in an existing IT
landscape with ArchiMate. However, we discovered multiple shortcomings:
1. The user is not guided through the modeling process, i.e., a formal modeling
procedure as for example used in [
For example, it may not be clear if a node element represents a speci c node of the network or a type of node. While it is not practical to represent all instances of nodes in the public Ethereum peer-to-peer network, the participating nodes for the particular application need to be modeled in Ethereum as well as in Hyperledger Fabric, where the network is managed. 4. The ArchiMate language lacks concepts speci c to DLT and therefore presents ambiguities in the available relationships and elements. For example, ow and trigger relations may or may not imply an ordering of process steps in the way shown for DLT consensus. 5. The structure of an application is not clearly separated from its behavior.
For example, application components and their interactions as an order of calling speci c functions are shown in one model. It might be desirable to specify the behavior based on the structure of the speci ed application. 6
In this paper we conducted a preliminary evaluation of the application of an enterprise modeling language to the domain of blockchains. For this purpose we created an ArchiMate model of a use case for NFT tokens that has been prototypically realized using Ethereum and Hyperledger. The preliminary analysis showed that the concepts currently contained in ArchiMate do not permit a fully adequate representation of blockchain-based applications. Therefore, it will be investigated in future work, which speci c extensions are necessary to enhance the adequacy and o er better modeling support for blockchain-based systems. In particular, we will consider extensions and pro les for ArchiMate for extending the scope of the language as well as the addition of formal modeling procedures.
The research on this paper has been partially Science Fund grant number 196889. nanced by the Swiss National