Enhancing Edge Node Resilience through SDN-Driven Proactive Failure Management
Sowmya R, Nandhini M, M. Priyanga · 2024
With the rapid advancements in technology, there has been a significant shift in how data is processed. We have moved from the traditional monolithic server setups to cloud computing, and now we are embracing the Internet of Things (IoT). With the growing prevalence of smart devices and real-time applications, the demand for edge computing has become essential to achieve real-time data processing with minimal latency and improved response times. Extensive research is currently underway in the field of edge computing, primarily focusing on reducing latency, enhancing response times, optimizing caching strategies, and utilizing heuristic algorithms for caching placement and replacement. The combination of Software-Defined Networking (SDN) with edge computing further augments the system's capabilities. Most research in SDN-enhanced edge computing has centered on topics like Service Function Chain (SFC) placements and caching models. Though all the above-mentioned approaches play a major role in edge computing, there has been limited attention given to ensuring the availability of edge servers or nodes. While some research has explored failure recovery for edge server applications, there has not been a significant emphasis on physical availability. This paper aims to bring spotlight on the importance of edge server availability, both at the system level and the application level. It proposes an approach that introduces a management plane responsible for detecting edge server failures and redeploying applications when an edge application crashes. Additionally, it leverages the SDN control plane to reroute traffic to a pre-configured backup edge server in the event of a system breakdown.