Functions as a Service (FaaS) is a crucial element of cloud computing, particularly for machine learning inferences. To reduce latency, cloud offloading has been proposed, positioning the FaaS provider closer to the client on the Edge-Cloud continuum. We extend this approach by proposing a multiprovider, fully decentralised model to enhance the flexibility, scalability, and resilience of FaaS. Additionally, we identify the necessary security properties and available countermeasures.