Internet of Things (IoT) has gained substantial attention over the past years. And the main discussion has been how to process the amount of data that it generates which has lead to the edge computing paradigm. Wether it is called fog1, edge or mist, the principle remains that cloud services must become available closer to clients. This documents presents ongoing work on future edge systems that are built to provide steadfast IoT services to users by bringing storage and processing power closer to peripheral parts of networks. Designing such infrastructures is becoming much more challenging as the number of IoT devices keeps growing. Production grade deployments have to meet very high performance requirements, and end-to-end solutions involve significant investments. In this paper, we aim at providing a solution to extend the range of the edge model to the very farthest nodes in the network. Specifically, we focus on providing reliable storage and computation capabilities immediately on wireless IoT sensor nodes. This extended edge model will allow end users to manage their IoT ecosystem without forcibly relying on gateways or Internet provider solutions. In this document, we introduce Achlys, a prototype implementation of an edge node that is a concrete port of the Lasp programming library on the GRiSP Erlang embedded system. This way, we aim at addressing the need for a general purpose edge that is both resilient and consistent in terms of storage and network. Finally, we study example use cases that could take advantage of integrating the Achlys framework and discuss future work for the latter.