Abstract. The complex distributed systems of nowadays require the dynamic composition of multiple components, which are autonomous and so complex that they can be considered as systems in themselves. These components often use different application protocols and are implemented on top of heterogeneous middleware, which hamper their successful inter-action. The explicit and rigorous description and analysis of components interaction is essential in order to enable the dynamic composition of these components. In this paper, we propose a formal approach to rep-resent and reason about interactions between components using layered connectors. Layered connectors describe components interaction at both the application and middleware layers and make explicit the role of mid-dleware in the realisation of this interaction. We provide formal semantics of layered connectors and present an approach for the synthesis of layered connectors in order to enable the dynamic composition of highly hetero-geneous components. We validate our approach through a case study in the area of collaborative emergency management.

Security is concerned with the protection of assets from in-tentional harm. Secure systems provide capabilities that en-able such protection to satisfy some security requirements. In a world increasingly populated with mobile and ubiqui-tous computing technology, the scope and boundary of se-curity systems can be uncertain and can change. A single functional component, or even multiple components indi-vidually, are often insufficient to satisfy complex security requirements on their own. Adaptive security aims to enable systems to vary their protection in the face of changes in their operational envi-ronment. Collaborative security, which we propose in this paper, aims to exploit the selection and deployment of mul-tiple, potentially heterogeneous, software-intensive compo-nents to collaborate in order to meet security requirements in the face of changes in the environment, changes in assets under protection and their values, and the discovery of new threats and vulnerabilities. However, the components that need to collaborate may not have been designed and implemented to interact with one another collaboratively. To address this, we propose a novel framework for collaborative security that combines adaptive security, collaborative adaptation and an explicit representation of the capabilities of the software components that may be needed in order to achieve collaborative secu-rity. We elaborate on each of these framework elements, focusing in particular on the challenges and opportunities afforded by (1) the ability to capture, represent, and reason about the capabilities of different software components and their operational context, and (2) the ability of components to be selected and mediated at runtime in order to satisfy the security requirements. We illustrate our vision through a collaborative robotic implementation, and suggest some areas for future work.