USA: Faster update for SDN-based internet of things sensory environments

Abstract

Internet of Things (IoT) require ubiquitous and seamless network connectivity. Meanwhile, it also asks for effective service management, data transmission and analysis. Towards this end, software defined networks (SDN) technology is introduced as a key solution to enable IoT network management. When IoT requires an efficient forwarding policy reconfiguration as the response to the alteration of system requirement (e.g., network environment or user demand), SDN is able to adjust the configurations easily to meet its demand. Previous research efforts tried to complete the reconfiguration as quickly as possible, since the update speed is critical to the performance of network management. However, update time can be further reduced. In this paper, we propose a novel update mechanism, which is referred to as rule co-existence update. It is able to enlarge the solution space and obtain more individual solutions, without any negative effect to the packet headers or throughput of flows. Then, we propose a scheme called USA (i.e., Update Speed Accelerator), as a novel algorithm for network updates based on the above-mentioned rule co-existence, that accelerates the update of forwarding policy with spare rule space on current network switches for general IoT environments. Based on the obtained suboptimal greedy solution, USA shows a novel way for rule co-existence to accelerate the update of forwarding policies. It also proposes a simplified algorithm to ensure the forwarding correctness during the update process. Extensive simulation results show that USA can clearly reduce the update time for nearly half of policies in our experiments with few rule space overhead, which is less than 8% of all policies, and maximum update time is less than previous works when spare rule space is limited.