Auctions for Secure Multi-party Policy Negotiation in Ambient Intelligence

The advent of distributed and ad-hoc-connected systems such as in Ambient Intelligence applications confronts developers with the question on how to specify QoS- and security policies beforehand, without knowing the exact capabilities and requirements of the platforms which will be present at run time. Especially when peers need to adhere to an appropriate trade-off between security and performance, on-the-fly negotiation protocols are required to allow peers to autonomously agree on a common policy. In this paper we present a framework for secure multi-party decision protocols based on auctions. Besides the framework design, a prototype implementation will be described and results of practical experiments will be given.