Biological networks, such as that of the human brain, show a remarkable ability to adapt to changing environments by long-term evolution and short-term adaptations. This is facilitated by the core-periphery structure of the network, where a core of densely connected network nodes provides long-term functionality and robustness, while the periphery is responsible for adaptation on short time scales to changing conditions in the environment. In this paper, we discuss the characteristics of the core-periphery network structure and its relation with adaptability and evolvability, which we also illustrate with own experimental data from the brain’s functional network. Based on this concept, we propose a mechanism for the adaptive placement of network functions to virtual servers in network function virtualization (NFV) under time-varying user requests. Our simulation results show that the number of server manipulations (migrations, merges, and replications of virtual machines) necessary to accommodate changes in network function requests can be greatly reduced compared to the conventional placement method.