Narrowband Internet of Things (NB-IoT) is a promising cellular technology for enabling low cost, low power, long-range connectivity to the IoT devices. With the bandwidth requirement of 180 kHz, it provides the flexibility to deploy within the existing Long Term Evolution (LTE) band. However, this raises serious concerns about the performance of the technology due to severe interference from multi-tier 5G heterogeneous networks (HetNets). Furthermore, as NB-IoT is based on halfduplex frequency division duplexing (HD-FDD), the symmetric allocation of spectrum band between downlink and uplink results in underutilization of resource, particularly in case of asymmetric traffic distribution. Therefore, an innovative radio resource management (RRM) strategy needs to be devised to improve the spectrum efficiency and device connectivity. This article presents the detailed design challenges that need to be addressed for the RRM of NB-IoT and proposes a novel framework to devise an efficient resource allocation scheme by exploiting cooperative interference prediction and flexible duplexing techniques.