Exploiting Concurrency for Opportunistic Forwarding in Duty-Cycled IoT Networks
Exploiting Concurrency for Opportunistic Forwarding in Duty-Cycled IoT Networks
复制标题
利用并发性实现占空比物联网网络中的机会转发
DOI:
10.1145/3322496
复制
发表时间:
2019
影响因子:
4.1
通讯作者:
Jiang Hongbo
中科院分区:
文献类型:
--
作者:
Liu Daibo;Cao Zhichao;He Yuan;Ji Xiaoyu;Hou Mengshu;Jiang Hongbo
Due to limited energy supply of Internet of Things (Zhao et al. 2018) (IoT) devices, asynchronous duty cycle radio management is widely adopted to save energy. Since the sleep schedules of nodes are unsynchronized, a sender has to repeatedly send frames to coordinate with its receiver or keep sleeping until the receiver’s wake-up time will come according to receiver’s sleep-wake schedule. In such contexts, opportunistic forwarding, which takes the earliest forwarding opportunity instead of a deterministic forwarder, shows great advantage in utilizing channel resource for duty-cycled IoT networks. The multiple forwarding choices with temporal and spatial diversity increase the chance of collision tolerance in opportunistic forwarding, potentially enhancing the overall performance of duty-cycled multi-hop networks. However, since the current channel contention mechanisms mainly focus on collision avoidance, it is too conservative to exploit concurrency. To address this problem, in this article, we propose COF to fully exploit the potential Concurrency for Opportunistic Forwarding in duty-cycled IoT networks. COF achieves concurrent transmission by: (i) measuring conditional link quality under the interference of on-going transmissions, and then (ii) further modeling the benefit of potential concurrency opportunities. According to the expected benefit of concurrency, COF decides whether or not to transmit in concurrent way. COF also adopts concurrency flag and signal features to avoid data collision caused by disordered concurrent transmissions and enhance the accuracy of conditional link quality estimation. COF can be easily integrated into the conventional unsynchronized and duty-cycled protocols. We have implemented COF and evaluated its performance on a 40-node testbed. The results show that COF can effectively exploit potential concurrency in opportunistic forwarding and COF outperforms the state-of-art protocols under diverse traffic load and network density.