Scientists:	Sree Harsha Totakura, M. Sc, Dr. Heiko Niedermayer, Prof. Dr.-Ing. Georg Carle
Duration:	01.09.2015 – 30.09.2018
Funding:	EU Horizon 2020
Homepage:	http://www.safecloud-project.eu/

Description

Cloud infrastructures raise concerns regarding privacy, integrity, and security of offsite data. These concerns are addressed by encrypting the data to be stored in the cloud. However, if the data is encrypted, the cloud infrastructure can only be used as a backup for the data, but not for running computations on the data. This prevents us from using the computational capabilities of cloud infrastrutures.

To be able to use the computation capabilities and still provide privacy, integrity and security of the data in cloud infrastrutures, a specialized set of algorithms and cryptography are needed. We find these in the fields of Secure Multi-party Computations (SMC), Homomorphic Encryption, and Erasure Resistant Encodings.

Secure Multi-party Computation algorithms allow two parties to compute a result without either party knowing the inputs of the other party. This helps to preserve the privacy of the involved parties as their data is kept private throughout the computation. Any information that is given out about their input is only through the result acquired from the computed function. Obviously, not every function could be computed in this way. However, the current state-of-the-art allows us to compute some functions which are already useful for practical use-cases, albeit moderate to high computational costs. SafeCloud project aims to realize such a use-case by working with a healthcare systems provider. Furthermore, it explores practical implementations of new algorithms to reduce the involved computational costs.

Homomorphic Encryption is used to compute functions on encrypted data. A subset of SMC functions can be realized using this type of encryption. The results of such computations may be encrypted and can only be known to the parties providing the encrypted inputs to the function. This allows us to use the cloud infrastructure for computation while keeping the data encrypted.

Erasure Resistance Encodings are helpful to tangle data from a customer with that of other customers. The tangled data cannot be deleted without severely corrupting the data it is tangled with. This is useful to provide service guarantees by a service provider because the service provider, or an attacker, has no way to delete the data of a customer without deleting data of other customers.

Our contribution to the project is in the development of secure communications middleware. Together with INESC-ID, Portugal we explore ways to provide vulnerability-tolerant communication channels, protected service provising, route monitoring, and multi-path communications.

Partners:

• Cloud&Heat, Germany
• Cybernetica, Estonia
• INESC-ID, Portugal
• INESC-TEC, Portugal
• Maxdata Software, Portugal
• Université de Neuchâtel, Switzerland
• Technische Universität München, Germany