Exploring heterogeneous ressources in cloud environments Authors: Max Plauth To date, the prevalent CPU architecture employed in cloud-based environments is x86_64. However, competing architectures such as ARM or Power8 are gaining momentum in servers. But not only the CPU landscape faces the potential of becoming more diverse, as hardware accelerators such as GPUs have become increasingly popular in many domains. Even though several cloud providers offer virtual machines with attached accelerator-resources, the underlying virtualization mechanisms lag behind the qualities of CPU-based virtualization. The goal of my work is to explore strategies for incorporating heterogenous compute resources whist taking into account the capabilities of different CPU architectures.