The ability to incorporate intelligence in even small devices and to make use of contextual information from widely deployed sensors has already begun to change management paradigms. As edge computing and IoT become more prevalent, systems will increasingly consist of cooperating, heterogeneous, distributed, autonomous elements. Architectures for cognitive, collaborative systems are evolving to deal with such complex environments. Concepts from multi-agent systems and autonomic computing are being applied to cope with the scope and breadth of large collections of interacting devices and services. Technologies for security and access control must evolve as well. Policy-based mechanisms are widely used and have been very successful in protecting information and controlling access to systems and services. They tend to rely, however, on a centralized infrastructure and on the automated enforcement of directives. Newer paradigms are being investigated that allow policy structures to be more dynamic and contextual, while still preserving the desired levels of control. We will present trends in the evolution of architectures for distributed, federated systems, and the technologies for managing them.