On Storing Data Objects of Business Pro...

Blockchain systems provide strong support for enforceability if all data relevant for transaction verification is stored on-chain. An example is the effective prevention of double spending in the bitcoin network. However, this approach has substantial drawbacks if privacy is a concern. One alternative in the enterprise context are architectures where access to the blockchain can be restricted to specific participants, and separate blockchains (channels) between subsets of participants can be established. While such architectures are promising for supporting privacy, there can still be trade-offs between the supported levels of enforcement and the degree of privacy that can be natively supported by those systems. In this paper, we follow a model-based approach, where privacy and enforceability requirements are explicitly modeled in business processes. We provide a detailed analysis of how different levels of data privacy affect the possible placement of data in channels on Hyperledger Fabric (HLF) and to what degree decisions over data with privacy requirements can be enforced by the built-in transaction verification mechanism. Finally, we present a prototype based on distributed oracles that allows us to enforce the correctness of decisions over data located on different channels.