Named functions and cached computations

Current ICN research favors a key-value-store view of the network, where location agnostic names typically resolve to documents, data blocks or sensor values. We believe that names should not only refer to data but also to functions and computation tasks. In Named Function Networking (NFN) the network´s role becomes to resolve names to computations, par example by reducing λ-expressions. In doing so, the network starts acting like a computing machine, capable of not only caching content but also computation results. We present basic concepts of NFN and report on our implementation that embeds the name resolution logic of CCNx in a generic resolver of λ-expressions. We demonstrate its resolution power beyond mere content-pull, to also leverage code-drag and computation-push as well as generalizing CCNx protocol functions.