[Paper Review] Multiple Access Computational Offloading: Communication Resource Allocation in the Two-User Case (Extended Version)

Mahsa Salmani, Timothy N. Davidson|arXiv (Cornell University)|May 14, 2018

IoT and Edge/Fog Computing21 references12 citations

TL;DR

This paper proposes an optimal communication resource allocation framework for two-user mobile edge computing systems with computational offloading, analyzing energy minimization under various multiple access schemes. It derives closed-form and quasi-closed-form solutions for indivisible and infinitesimally partitionable tasks, showing that full multiple access channel utilization significantly reduces energy compared to TDMA when channel gains are asymmetric and latency constraints are tight.

ABSTRACT

By offering shared computational facilities to which mobile devices can offload their computational tasks, the mobile edge computing framework is expanding the scope of applications that can be provided on resource-constrained devices. When multiple devices seek to use such a facility simultaneously, both the available computational resources and the available communication resources need to be appropriately allocated. In this manuscript, we seek insight into the impact of the choice of the multiple access scheme by developing solutions to the mobile energy minimization problem in the two-user case with plentiful shared computational resources. In that setting, the allocation of communication resources is constrained by the latency constraints of the applications, the computational capabilities and the transmission power constraints of the devices, and the achievable rate region of the chosen multiple access scheme. For both indivisible tasks and the limiting case of tasks that can be infinitesimally partitioned, we provide a closed-form and quasi-closed-form solution, respectively, for systems that can exploit the full capabilities of the multiple access channel, and for systems based on time-division multiple access (TDMA). For indivisible tasks, we also provide quasi-closed-form solutions for systems that employ sequential decoding without time sharing or independent decoding. Analyses of our results show that when the channel gains are equal and the transmission power budgets are larger than a threshold, TDMA (and the suboptimal multiple access schemes that we have considered) can achieve an optimal solution. However, when the channel gains of each user are significantly different and the latency constraints are tight, systems that take advantage of the full capabilities of the multiple access channel can substantially reduce the energy required to offload.

Motivation & Objective

• To develop a centralized resource allocation strategy that minimizes mobile device energy consumption in multi-user computational offloading.
• To analyze the impact of multiple access schemes—especially full-channel-capability utilization versus TDMA—on energy efficiency in offloading.
• To derive closed-form and quasi-closed-form solutions for energy minimization under indivisible and partitionable tasks.
• To evaluate the performance gains of advanced multiple access techniques over conventional schemes like TDMA under asymmetric channel conditions.
• To provide design insights for practical offloading systems by characterizing the trade-offs between latency, power, and channel utilization.

Proposed method

• Formulates a joint communication and computational resource allocation problem for two mobile users offloading to a shared edge server.
• Derives closed-form solutions for energy minimization under full multiple access channel utilization and TDMA for indivisible tasks.
• Uses coordinate descent algorithms to solve quasi-closed-form problems for partitionable tasks and mixed task types.
• Incorporates constraints from latency requirements, transmission power limits, and achievable rate regions of different multiple access schemes.
• Applies sequential decoding and independent decoding models to analyze suboptimal schemes in comparison to optimal full-capability systems.
• Reduces dimensionality of optimization problems using equality constraints from active latency constraints at optimality.

Experimental results

Research questions

• RQ1 How does the choice of multiple access scheme affect the energy efficiency of mobile computational offloading in a two-user system?
• RQ2 Under what channel and power conditions does TDMA achieve optimal energy performance compared to full-capability multiple access?
• RQ3 What are the closed-form or quasi-closed-form solutions for energy-minimizing resource allocation in the presence of indivisible and partitionable tasks?
• RQ4 How do asymmetric channel gains and tight latency constraints influence the performance gap between advanced multiple access and conventional schemes?
• RQ5 Can equality constraints from active latency constraints be leveraged to reduce the dimensionality of the optimization problem?

Key findings

• When channel gains are equal and transmission power exceeds a threshold, TDMA achieves optimal energy performance.
• Systems exploiting the full capabilities of the multiple access channel achieve significantly lower energy consumption than TDMA when channel gains are asymmetric and latency constraints are tight.
• For indivisible tasks, closed-form solutions are derived for full-capability and TDMA schemes, while quasi-closed-form solutions are obtained for sequential and independent decoding.
• The optimization problem for mixed task types (indivisible and divisible) is reduced in dimension by leveraging equality constraints from active latency constraints.
• The objective function in the reduced problem remains quasi-convex in each variable, enabling convergence via coordinate descent algorithms.
• The derived solutions demonstrate that advanced multiple access techniques can substantially reduce mobile energy consumption in asymmetric channel environments.