Honeybees are essential to human society, providing pollination services globally as well as producing honey and other valuable products. Effective management of apiaries should not only rely on beekeeper knowledge and skill, but also incorporate new information technologies. The options to identify, predict and prevent beekeeping problems are becoming more affordable and applicable. The interdisciplinary Horizon 2020 project HIVEOPOLIS focuses on developing a new approach in beekeeping, by creating novel mechatronic beehives and implementing new bio-hybrid ideas. These intelligent beehives aim to help honeybees to cope with adverse environmental factors and increase the survival rate of the bee colonies. This paper focuses on the software architecture design for these intelligent beehives, providing infrastructure for data management and decision support system operation. The presented infrastructure is suitable for highly dynamic and diverse environments where a multitude of components interact and exchange information across technology domains (embedded, cloud, UIs) in a reliable and secure way. Besides user support, the decision support system built upon this infrastructure enables closed-loop automated decision making and control.