Work

Aspect-Oriented Approach of Non-Functional Requirements for Embedded System-Based IoT Applications

Year: 2025

Type: article

Abstract: Designing high-performance embedded devices is usually time consuming and error-prone due to conflicting non-functional requirements (NFRs) like performance, timeliness, and sustainability. This paper proposes an Aspect-Oriented Programming (AOP) approach using AspectJ to streamline IoT middleware design by modularizing NFRs and communicating domain knowledge to optimizers and mapping tools. Leveraging AspectJ, our method enhances modularity by separating crosscutting concerns—such as scheduling, resource allocation, and synchronization—from core functionality. Our novel contribution lies in refactoring crosscutting concerns— task scheduling, resource management, and task synchronization —into reusable aspects, achieving a Cohesion Index (CoI) increase from 0.19 (Java) to 0.69 (AspectJ), a 263% improvement, while reducing code tangling. By leveraging context primitives and behavior functions aligned with standards like OSEK OS, we address IoT challenges such as real-time criticality and resource scarcity Validated across a real-time audio switch and simulated FreeRTOS environments, our method outperforms traditional Java designs, cutting memory usage by 20% (24.25% to 20.01%) and CPU utilization by 9% (44% to 40%), with a 50% runtime reduction in critical tasks (e.g., 92.03μs to 46.19μs). Unlike prior AOP efforts focused on desktop systems, this approach targets embedded IoT, offering a scalable, standards-integrated solution. A survey of real-time embedded system development highlights the superiority of Aspect-Oriented Software Development (ASOD) in enhancing schedulers, resource allocation agents, and other synchronization mechanisms with minimal overhead, setting a foundation for broader IoT adoption.

Source: IEEE Access

DOI: 10.1109/access.2025.3585417

Authors: Anirban Bhar, Sabnam Sengupta, Sankhayan Choudhury

Institution: University of Calcutta

Language: English

Topic: Advanced Software Engineering Methodologies

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

SDG: Industry, innovation and infrastructure

Open Access status: gold

Estimated APC paid: $2,075