3835

Business processes may span multiple organizations. For instance, cattle may move through several organizations in an agricultural supply chain, or patients may be seen by multiple healthcare providers as part of a treatment process. Optimizing these cross-organizational processes is an aim of process mining, however, process mining efforts may be challenged by commercially sensitive data or privacy laws, which may prevent the involved organizations from sharing recorded process data with one another. Federated process mining aims to perform inter-organizational analyses without information being shared across organizational borders. In this paper, we propose a federated technique to discover directly-follows graphs (DFGs), using homomorphic encryption, while keeping timestamps and activities secret. We evaluate the feasibility of our technique using real-life event logs to discover their DFGs and discuss potential attacks.