Integration as a service (INTaaS) is the centrepiece of current corporate, cloud and device integration processes. Thereby, compositions of integration patterns denote the required integration logic as integration processes, currently running in single-clouds. While multicloud settings gain importance, their promised freedom of selecting the best option for a specific problem is currently not realized as well as security constraints are handled in a cost-intensive manner for the INTaaS vendors, leading to security vs. costs goal conflicts, and intransparent to the process modeler. In this work, we propose a design-time placement for processes in multiclouds that is cost-optimal for INTaaS problem sizes, and respects configurable security constraints of their customers. To make the solution tractable for larger, productive INTaaS processes, it is relaxed by using a local search heuristic, and complemented by correctness-preserving model decomposition. This allows for a novel perspective on cost-aware process modeling from a process modeler's perspective. The multicloud process placement is evaluated on real-world integration processes with respect to cost- and runtime-efficiency, and discusses interesting trade-offs. The process modeler's perspective is investigated based on a new cost-aware modeling process, featuring the interaction between the user and the INTaaS vendor through ad-hoc multicloud cost calculation and correctness-preserving, process cost reduction proposals.
View on arXiv