Edge computing represents a combination of distributed computing connected to centralized servers. Actors on the Edge may interact with each other as well as a central data center. Their concerns include multiple subtopics, e.g., protecting information content from observation and alteration, protection of operational capability from unauthorized access, protection of normal operation in the presence of malicious overloaded requests, etc. Therefore, trust requires a distributed solution. In centralized learning, the central server potentially represents a single point of failure, which is one of the bottlenecks for performance as well. Another issue is the need for all participants to trust the central authority with their datasets. In contrast, a decentralized federated learning solution needs parties to run a common binary on each of their datasets and trust the incoming program, thus avoiding a single point of failure but potentially creating a security hazard with malicious code. Another issue is the long training run time due to multiple hops between different dataset locations. In this chapter, we propose a novel collaborative federated learning (CFL) solution that combines the advantages of centralized and decentralized federated schemes without compromising security.