In today's dynamic business scenarios, organizations cooperate to achieve their business goals. These cooperations are modeled using choreographies, specifying the externally visible interaction behavior of each participant. Thus, the choreography represents a shared behavioral contract, where each participant designs its internal process in a way that matches the choreography. Existing choreography literature and standard modeling notations generally assumes that all participants know and trust each other.In this project we relax this assumption and explore how to execute choreographies among a dynamic set of mutually distrustful participants. In this setting, it is not possible to rely on a centrally managed process execution log, so blockchain technology provide a valid solution. This project will establish a new research collaboration between USI and HPI to research concepts, methods and techniques for studying the impact of emerging blockchain technologies on flexible choreography design as well as how choreography execution can provide a unique use case scenario for multi-chain interoperability.The state of the art in choreography modeling presumes that all interactions are known before the choreography execution starts. This static view is not adequate to support cases in which neither all interaction partners nor all interactions needed to achieve a business goal can be known at design time. This project will explore different fragment-based techniques and policies to introduce flexibility in choreography design and execution. Blockchains are a powerful coordination medium to build a shared view over a log of interaction events among a network of untrusted partners. In this project we will use the blockchain primitives to support the execution of choreographies, which give a global representation of how a set of independent participants interact. In particular, we will provide a mapping of choreography fragments to smart contracts. We will show how smart contracts, the essential building block of blockchain-based applications, can be generated from choreography models and used to enforce the choreography model constraints and execute the corresponding interactions through the blockchain. To do so we will need to find a suitable level of granularity for the mapping between models and smart contract code. The project will go beyond current limitations of blockchain-based business process execution in which a single blockchain stores the process instances execution log. Due to real-world constraints, whereby geographical and jurisdictional boundaries play a major role concerning which blockchain network is accessible to different participants, in this project we will study how ongoing standardization efforts to achieve blockchain interoperability and smart contract portability can be leveraged for flexible choreography execution across multiple blockchains.
Submitted by cp on