On the Efficiency of Job Offloading in Edge Networks

One of the main challenges of edge computing is to establish efficient mechanisms for computing and offloading jobs in edge systems, such that their computation time and latency are minimized. This paper presents a simulation-based study of job offloading in edge computing scenarios. Our findings indicate the importance of using optimal parameters and offloading strategies to achieve efficient computing in terms of computation times and communication overhead, even with relatively simple probabilistic job offloading schemes. Consequently, job offloading operating with a suboptimal configuration can congest the entire edge network and lead to an increased computation time. Ultimately, such suboptimal configurations can lead to results that are so inferior, that they contradict the rationale of using edge computing for latency critical applications.