Server Placement for Edge Computing: A Robust Submodular Maximization Approach

In this work, we study the problem of Robust Server Placement (RSP) for edge computing, i.e., in the presence of uncertain edge server failures, how to determine a server placement strategy to maximize the expected overall workload that can be served by edge servers. We mathematically formulate the RSP problem in the form of robust max-min optimization, derived from two consequentially equivalent transformations of the problem that does not consider robustness and followed by a robust conversion. RSP is challenging to solve, because the explicit expression of the objective function in RSP is hard to obtain, and it is a robust max-min problem with knapsack constraints, which is still an unexplored problem in the literature. We reveal that the objective function is monotone submodular, and propose two solutions to RSP. First, after proving that the involved constraints form a $p$p-independence system constraint, where $p$p is a parameter determined by the coefficients in the knapsack constraints, we propose an algorithm that achieves a provable approximation ratio in polynomial time. Second, we prove that one of the knapsack constraints is a matroid contraint, and propose another polynomial time algorithm with a better approximation ratio. Furthermore, we discuss the applicability of the aforementioned algorithms to the case with an additional server number constraint. Both synthetic and trace-driven simulation results show that, given any maximum number of server failures, our proposed algorithms outperform four state-of-the-art algorithms and approaches the optimal solution, which applies exhaustive exponential searches, while the proposed latter algorithm brings extra performance gains compared with the former one.

Automated summary

In this work, the problem of Robust Server Placement (RSP) for edge computing is studied. The objective is to determine a server placement strategy that maximizes the expected overall workload that can be served by edge servers in the presence of uncertain server failures. The RSP problem is formulated as a robust max-min optimization problem with knapsack constraints, which is challenging to solve. Two algorithms are proposed to solve the RSP problem, both achieving provable approximation ratios in polynomial time. Synthetic and trace-driven simulation results demonstrate the superiority of the proposed algorithms over state-of-the-art approaches.