Collaborative Research Centre 901 On-The-Fly Computing

Link to the official web page of CRC 901 On-The-Fly Computing

The objective of CRC 901 – On-The-Fly Computing (OTF Computing) – is to develop techniques and processes for automatic on-the-fly configuration and provision of individual IT services out of base services. The base serves are available on world-wide markets. In addition to the configuration by special OTF service providers and the provision by what are called OTF Compute Centers, this involves developing methods for quality assurance and the protection of participating clients and providers, methods for the target-oriented further development of markets, and methods to support the interaction of the participants in dynamically changing markets.

Successful marketing and acceptance of On-The-Fly Computing systems require dealing with their security. The highly dynamic and heterogeneous nature of the visioned system as well as data protection and further legal requirements pose a challenge for modern cryptography and require novel cryptographic solutions.

During the course of the project we intend to design solutions for confidential and authenticated communication in dynamic groups based on identity-based cryptography. This requires key revocation and reduction of power of authority for identity-based schemes. Efficient allocation of services and data access control in On-The-Fly Computing systems also require novel cryptographic schemes, which we intend to realize based on the attribute-based encryption schemes.

Another important market mechanism in On-The-Fly Computing Systems is an anonymous reputation system which enables clients to rate products and services and gives incentives to providers to improve their services. Hence, in this part of the project we develop new models of security and schemes to build a highly flexible and secure reputation system.

One of our current goals is to develop secure, efficient and flexible access control systems for On-The-Fly Computing Data Centers and On-The-Fly service providers. A promising approach is to use attribute-based encryption schemes. These novel schemes use techniques from pairing-based cryptography and are related to identity-based cryptography too.

In the ciphertext-policy attribute-based encryption schemes, the owner of data defines an access policy for each data and encrypts it once using this policy. The policies for different data are Boolean formulas over predefined attributes. In order to provide access to the encrypted data, the owner of data gets each customer with a special decryption key. Every key is related to a set of attributes. A customer will be able to decrypt a ciphertext if and only if the attributes of his/her key satisfy the policy of the ciphertext.

The attribute-based approach simplifies the realization of data access control systems, which then can be even stored on an untrusted server. The data access control is completely realized by the encryption and all the data must be encrypted only once for all the customers.

In this area, our research focus is on the development of efficient and flexible attribute-based encryption schemes. The policies of the existing schemes are restricted to several classes of functions and are quite inefficient. On the one hand, we are interested in the development of schemes which can be applied to general function classes. On the other hand, we are looking for more efficient methods when realizing restricted function classes. Other modifications of the schemes will be also necessary when considering further questions arising from privacy protection and further legal requirements.

A second goal is to develop anonymous reputation systems. To provide trustworthy, reliable, and honest ratings there is a need for anonymous reputation systems that also guarantee that customers rate products only once. To further increase trust in the system, everyone – even outsiders – should be able to verify the validity of ratings. Some of these properties have been studied in the context of group signatures. However, the concept of group signatures does not meet all the requirements for reputation systems. In particular, reputation systems do not consist of a single group of users. Rather one can think of reputation systems as a family of group signature schemes – one for each product. Moreover, we may have providers with several products. Hence, when looking at security and anonymity group signature schemes for different products can not be considered in isolation. Finally, known constructions of group signatures do not provide all properties that we need for a secure and anonymous reputation system and do not provide them simultaneously.

The research focus in the area of reputation systems is the development of new security models and efficient, flexible and secure schemes which meet all our requirements. Here we mainly consider group signatures, but also attribute-based signatures and anonymous credential systems will be taken into account.