Emerging network scenarios call for innovative open service frameworks to ensure capability of self-adaptability and long-lasting evolvability. In this paper, we assess the need for such innovative service frameworks, and discuss how their engineering should get inspiration from natural ecosystems, i.e., by modelling services as autonomous individuals in an ecosystem of other services and data sources. We introduce a reference conceptual architecture with the goal of clarifying the concepts expressed and framing the several possible nature-inspired metaphors that could be adopted to realise the idea. On this basis, we go into details about one of such possible approaches, in which the rules governing the ecosystem are inspired by biochemical mechanisms. A case study is also introduced to exemplify the potentials of the presented biochemical approach and to experiment with some representative biochemistry-inspired patterns of adaptive service organisation and evolution.