Low-power Lossy Wireless Networks (LLWNs) are characterized by constraints in memory, processing, and power consumption, coupled with an inherently dynamic wireless environment. In this context, a suite of programmable communication protocols is essential to efficiently adapt to varying network conditions, optimize resource utilization, and maintain performance within the stringent limitations of LLWN devices. In this work, we review and compare state-of-the-art network programmability techniques to assess their suitability for LLWNs. Based on the findings, we propose a new network architecture for LLWNs, utilizing software-defined networking for control plane programmability and Femto Containers lightweight virtualization for data plane programmability, ensuring it respects the constraints of LLWN devices. We have conducted a proof-of-concept validation to demonstrate the feasibility of Femto Containers to implement the data plane in LLWN on the FIT IoT-LAB testbed. The results show that our architecture successfully achieves a substantial reduction in update size while adhering to memory and power consumption constraints of LLWN devices, although this comes at the cost of a slightly acceptable increased packet processing delay.