Deadline-constrained Multi-resource Task Mapping and Allocation for Edge-Cloud Systems

TL;DR

This paper addresses the complex problem of task offloading and resource allocation in edge-cloud systems with deadline constraints, proposing algorithms to optimize profit while meeting service deadlines.

Contribution

It formulates the task mapping and resource allocation as a non-convex MINLP problem and introduces two algorithms, ZSG and LDM, to effectively solve it.

Findings

ZSG performs within 3% of LDM with a 5-unit minimum.

ZSG outperforms LDM by 6.88% with a 15-unit minimum.

Algorithms improve resource utilization and deadline adherence.

Abstract

In an edge-cloud system, mobile devices can offload their computation intensive tasks to an edge or cloud server to guarantee the quality of service or satisfy task deadline requirements. However, it is challenging to determine where tasks should be offloaded and processed, and how much network and computation resources should be allocated to them, such that a system with limited resources can obtain a maximum profit while meeting the deadlines. A key challenge in this problem is that the network and computation resources could be allocated on different servers, since the server to which a task is offloaded (e.g., a server with an access point) may be different from the server on which the task is eventually processed. To address this challenge, we first formulate the task mapping and resource allocation problem as a non-convex Mixed-Integer Nonlinear Programming (MINLP) problem, known as NP-hard. We then propose a zero-slack based greedy algorithm (ZSG) and a linear discretization method (LDM) to solve this MINLP problem. Experiment results with various synthetic tasksets show that ZSG has an average of $2.98\%$ worse performance than LDM with a minimum unit of 5 but has an average of $6.88\%$ better performance than LDM with a minimum unit of 15.