Named Data Networking (NDN) changes the network communication paradigm from address-centric in TCP/IP to data-centric. This architectural change extends capabilities of the network's forwarding plane, enabling more efficient and adaptive data retrieval. However, it is unclear how to utilize forwarding behaviors, and reduce the deployment complexity of NDN at edge networks. In fact, to run NDN applications between just two nodes involves decent knowledge of the software forwarder (NFD) and non-trivial network configuration. This paper designs self-configured adaptive forwarding plane for NDN at edge networks, to improve data retrieval efficiency and minimize deployment complexity. The designed functionalities are implemented in NFD and tested on real devices. Results show that NFD is able to run out-of-the-box with different applications. More specifically, the new forwarding strategy is able to learn and utilize multiple faces and routes, and to better handle NACK and link failure, achieving good performance.