In this paper, we propose a Network Knowledge Layer (NKL) that is a service platform overlaid onto the existing networks to provide network knowledge for user-centric services in the Next Generation Network (NGN). Most traditional networks lack capabilities for accommodating user-centric service paradigm. Taking this into consideration, the proposed NKL has capabilities to acquire contextual information from various sources, to evolve this information with legacy information into high-level knowledge, and to expose the high-level knowledge to entities outside the network. For easy knowledge exposure, we specify a set of abstracted application programming interfaces (APIs). For efficient handling of network knowledge accessed by the APIs, we design and compare the three different network knowledge models. With the proposed APIs and the three knowledge models, we accomplish the experimental performance evaluation of NKL to verify its feasibility. The results of the various tests on knowledge models and APIs give good guidelines for efficient design and exposing network knowledge in developing a user-centric service platform. Finally, we expect NKL can support the effective development and execution of user-centric services by providing rich network knowledge with the APIs.