EAD-BRC: A Novel Edge Application Deployment Algorithm Balancing Robustness and Coverage

Application service providers are confronted with the challenge of effectively selecting server deployments amidst a vast array of nearby client-server clusters, all within constrained budgets, to ensure an optimal user experience. This challenge is encapsulated in the edge application deployment (EAD)problem. While existing research offers some viable solutions, two prevalent issues persist: the consideration of singular dimensions and the inability to achieve genuine balance across multiple dimensions due to differences in evaluation metric magnitudes. In response to the dilemma of balancing robustness and coverage in edge application deployment (BRCEAD) problem, we introduce two metrics of the same magnitude: robustness ratio and coverage ratio. We propose the use of the harmonic mean model $F1_{RC}$ to gauge their integrated benefits and develop the EAD-BRC algorithm with the objective of maximizing this comprehensive benefit. Comparative experiments with three baseline methods on a public dataset demonstrate that our approach not only effectively balances robustness and coverage but also excels in comprehensive benefits, while also performing admirably in single-dimensional benefits. This offers valuable insights into addressing the edge application deployment problem of balancing coverage and robustness.