학술논문

Edge Controller-Assisted SDN Architecture for Internet of Things
Document Type
Periodical
Source
IEEE Sensors Journal IEEE Sensors J. Sensors Journal, IEEE. 23(22):28182-28190 Nov, 2023
Subject
Signal Processing and Analysis
Communication, Networking and Broadcast Technologies
Components, Circuits, Devices and Systems
Robotics and Control Systems
Internet of Things
Computer architecture
Delays
Routing
Cloud computing
Sensors
Edge computing
Availability
edge computing
flow management
Internet of Things (IoT)
OpenFlow
scalability
software-defined network (SDN)
Language
ISSN
1530-437X
1558-1748
2379-9153
Abstract
Internet of Things (IoT) devices are resource and energy-constrained which forward many limitations such as packet loss, minimum storage capacity, lower computational ability, scalability, and availability. By introducing the concept of software-defined network (SDN) in IoT, a new variant namely SDN-based IoT has been designed. The traditional SDN-based IoT system uses a single centralized controller to manage the network. The use of a single controller can be convenient, however, it gives rise to issues such as single point of failure (SPoF). To overcome these limitations, we have proposed an edge-enabled SDN architecture namely ESDoT. The edge-enabled SDN IoT in ESDoT can provide a swift response to the data that require instant focus. The SDN controllers installed in the edge network can help to reduce the computation, communication, and management ability. In the ESDoT architecture, multiple SDN controllers consequently improve the scalability and availability feature by keeping the required services and flow rules at the edge. The ESDoT architecture has been evaluated using a testbed setup. The experimental results were compared with conventional SDN (CSDN)-based IoT and random selection of computational path for SDN-based IoT mode (RSCM). It was observed that ESDoT architecture can reduce the average round trip time (RTT) of the network by 18% and packet loss by 82%. The proposed architecture succeeded in reducing the end-to-end delay by 20% while increasing the system throughput by 32% when compared to the CSDN and RSCM.