LUTZ SCHUBERT, MATTHIAS ASSEL, ALEXANDER KIPP, and STEFAN WESNER
Outsourcing computation and/or storage away from the local infrastructure is not a new concept itself: Already the grid and Web service domain presented (and uses) concepts that allow integration of remote resource for seemingly local usage. Nonetheless, the introduction of the cloud concept via such providers as Amazon proved to be a much bigger success than, for example, Platform's Grid Support —or at least a much more visible success. However, the configuration and management overhead of grids greatly exceeds one of the well-known cloud providers and therefore encourages, in particular, average users to use the system. Furthermore, clouds address an essential economical factor, namely, elastic scaling according to need, thereby theoretically reducing unnecessary resource loads.
Cloud systems are thereby by no means introducing a new technology—just the opposite in fact, because many of the initial cloud providers simply opened their existing infrastructure to the customers and thus exploited their respective proprietary solutions. Implicitly, the offered services and hence the according API are specific to the service provider and can not be used in other environments. This, however, poses major issues for customers, as well as for future providers.
Interoperability and Vendor Lock-In. Since most cloud offerings are proprietary, customers adopting the according services ...