UBC Theses and Dissertations
Edge service placement and workload allocation under uncertainty Cheng, Jiaming
Edge computing (EC) has emerged as a promising architecture for hosting critical services with stringent latency and performance requirements, challenging to address in traditional cloud computing (CC) systems. EC makes distributed computation and storage resources close to end-users, providing low-latency and high-capacity services. Notable use cases of EC include real-time data analytic services, manufacturing automation, and computational offloading for the Internet of Things. Despite the tremendous potential, EC is still in its infancy stage, and many open problems remain to be solved. This thesis lies in the intersection of operations research and network economics, with a specific focus on developing mathematical models for decision-making and economic analysis of edge-cloud network systems. To support rapid response to incidents in EC, we propose a novel resilience-aware edge service placement and workload allocation model that jointly captures the uncertainty of resource demand and node failures. The salient feature of the proposed model identifies the optimal placement and procurement solutions that can hedge against all uncertain realizations of the traffic demand within an uncertainty set. Hence, it enables service providers to balance the trade-off between the operating cost and service quality while considering demand uncertainty and node failures. Furthermore, by leveraging the column-and-constraint generation (CCG) method, we introduce two iterative algorithms that can converge to an exact optimal solution within a finite number of iterations. We further suggest an affine decision rule (ADR) approximation approach for solving large-scale problem instances in a reasonable time. Extensive numerical results then demonstrate the advantages of the proposed model and solutions.
Item Citations and Data
Attribution-NonCommercial-NoDerivatives 4.0 International