Nowadays, an ever-increasing number of service
providers takes advantage of the cloud computing paradigm
in order to efficiently offer services to private users, businesses, and governments. However, while cloud computing
allows to transparently scale back-end functionality such as
computing and storage, the implied distributed sharing of
resources has severe implications when sensitive or otherwise
privacy-relevant data is concerned. These privacy implications
primarily stem from the in-transparency of the involved backend providers of a cloud-based service and their dedicated
data handling processes. Likewise, back-end providers cannot
determine the sensitivity of data that is stored or processed in
the cloud. Hence, they have no means to obey the underlying
privacy regulations and contracts automatically. As the cloud
computing paradigm further evolves towards federated cloud
environments, the envisioned integration of different cloud platforms adds yet another layer to the existing in-transparencies.
In this paper, we discuss initial ideas on how to overcome
these existing and dawning data handling in-transparencies and
the accompanying privacy concerns. To this end, we propose
to annotate data with sensitivity information as it leaves the
control boundaries of the data owner and travels through to
the cloud environment. This allows to signal privacy properties
across the layers of the cloud computing architecture and
enables the different stakeholders to react accordingly.