Pervasive computing calls for innovative open service frameworks that ensure capability of self-adaptability and long-lasting evolvability (i.e., eternity). One way to tackle this problem is by platforms enabling modelling and deployment of services acting just like autonomous individuals in an ecosystem of other services, data sources, and pervasive devices. Among the many nature-inspired paradigms that can be used to realise this idea, in this paper we focus on a biochemical approach. The application world is seen as a biological system (the net resembling a tissue of biological compartments like cells), and individuals are seen as chemical reactants floating in it: the underlying platform executes chemical laws which are responsible of system evolution in terms of production/decay of individuals, their spatial diffu- sion into the system, and their interaction. A computational model and some examples are shown to exemplify the ability of this framework to provide a self-organising ground to build eternally adaptive service ecosystems.