Optimal Edge User Allocation in Edge Computing with Variable Sized Vector Bin Packing (1904.05553v1)

Abstract: In mobile edge computing, edge servers are geographically distributed around base stations placed near end-users to provide highly accessible and efficient computing capacities and services. In the mobile edge computing environment, a service provider can deploy its service on hired edge servers to reduce end-to-end service delays experienced by its end-users allocated to those edge servers. An optimal deployment must maximize the number of allocated end-users and minimize the number of hired edge servers while ensuring the required quality of service for end-users. In this paper, we model the edge user allocation (EUA) problem as a bin packing problem, and introduce a novel, optimal approach to solving the EUA problem based on the Lexicographic Goal Programming technique. We have conducted three series of experiments to evaluate the proposed approach against two representative baseline approaches. Experimental results show that our approach significantly outperforms the other two approaches.

• The paper introduces a novel optimization model that formulates the edge user allocation problem as a variable sized vector bin packing problem using lexicographic goal programming.
• It demonstrates that the proposed method achieves near-optimal allocations and significantly reduces the number of edge servers compared to baseline methods.
• Experimental results across various scenarios highlight its scalability and practical implications for cost-efficient mobile edge computing deployments.

Optimal Edge User Allocation in Edge Computing with Variable Sized Vector Bin Packing

The paper addresses an optimization problem in the context of mobile edge computing (MEC), where efficient resource management is critical for enhancing service delivery. The authors focus on the Edge User Allocation (EUA) problem, a key challenge in MEC, characterized by the need to optimally allocate end-users to edge servers to balance the trade-offs between service accessibility and operational costs. The paper presents a novel approach modeling the EUA problem as a Variable Sized Vector Bin Packing (VSVBP) problem, an $\mathcal{NP}$-hard generalization of the classical Bin Packing (BP) problem.

Central to the paper is the utilization of Lexicographic Goal Programming (LGP) to address the EUA problem's dual objectives: maximizing the number of allocated users and minimizing the number of hired edge servers. Edge servers are distributed geographically, catering to end-users positioned within specific coverage areas. The optimization model developed considers these geographical constraints alongside multi-dimensional capacity constraints, represented by vector measures of computing resources such as CPU, memory, bandwidth, etc.

The evaluation of the proposed approach against random and greedy baseline methods is comprehensive. The authors conduct experiments varying the number of users, available edge servers, and remaining server capacities to simulate diverse real-world scenarios. The experimental results underscore the efficacy of the proposed method, highlighting its ability to achieve near-optimal allocations with a significantly reduced number of edge servers compared to baseline methods. Notably, the method sustains high efficacy as the problem scales, highlighting its practicality in large-scale MEC deployments.

The implications of this research are substantial for MEC frameworks, where optimal resource allocation can lead to significant reductions in operational costs associated with hiring edge servers. The innovative application of VSVBP and LGP techniques offers a robust framework that efficiently addresses the complex dynamics of MEC environments.

Moving forward, this research paves the way for tackling more dynamic and realistic scenarios where user mobility and demand variability are considered. Expanding the model to incorporate additional factors such as network latency, service availability, and economic considerations will be crucial for creating more comprehensive solutions to the EUA problem. This paper establishes a foundational approach that can be further refined and adapted to meet the evolving challenges in edge computing landscapes. Future developments may also explore adaptive algorithms that can respond in real-time to changes in user patterns and server availability, further enhancing the MEC's efficacy and reliability.