SD-NFV as an Energy Efficient Approach for M2M Networks Using Cloud-Based 6LoWPAN Testbed

摘要

Machine-to-Machine (M2M) communication is the leading technology for realising the Internet-of-Things (IoT). The M2M sensor nodes are characterized by low-power and low-data rates devices which have increased exponentially over the years. IPv6 over low power wireless personal area network (6LoWPAN) is the first protocol that provides IPv6 connectivity to the wireless M2M sensor nodes. Having a tremendous number of M2M sensor nodes execute independent control decision leads to difficulty in network control and management. In addition, these evergrowing devices generate massive traffic and cause energy scarcity, which affects the M2M sensor node lifetime. Recently, software defined-networking (SDN) and network functioning virtualization (NFV) are being used in M2M sensor networks to add programmability and flexibility features in order to adopt the exponential increment in wireless M2M traffic and enable network configuration even after deployment. This paper presents a proof-of-concept implementation which aims to analyze how SDN, NFV, and cloud computing can interact together in the 6LoWPAN gateway to provide simplicity and flexibility in network management. The proposed approach is called customized software defined-NFV (SD-NFV), and has been tested and verified by implementing a real-time 6LoWPAN testbed. The experimental results indicated that the SD-NFV approach reduced the network discovery time by 60% and extended the node’s lifetime by 65% in comparison to the traditional 6LoWPAN network. The implemented testbed has one sink which is the M2M 6LoWPAN gateway where the network coordinator and the SDN controller are executed. There are many possible ways to implement 6LoWPAN testbed but limited are based on open standards development boards (e.g., Arduino, Raspberry Pi, and Beagle Bones). In the current testbed, the Arduino board is chosen and the SDN controller is customized and written using C++ language to fit the 6LoWPAN network requirements. Finally, SDN and NFV have been envisioned as the most promising techniques to improve network programmability, simplicity, and management in cloud-based 6LoWPAN gateway.