Journal of Operating Systems Development & Trends · Vol. 13, No. 2 (2026)
Chip Interconnects, Smart Scheduling and Linux Malware: Journal of Operating Systems Development & Trends 13(2)
By STM Journals ·
This issue of the Journal of Operating Systems Development and Trends covers how computing systems are organised and protected. One paper proposes a new network-on-chip layout for 64-core processors. Another presents an AI-based scheduler for mixed CPU and GPU systems, and a third compares load-balancing methods for cloud computing. The remaining two review lightweight operating systems for low-power smart hardware and AI methods for detecting Linux malware.
What's in this issue
-
Mirror-Symmetric NoC Topologies for 64-Core Arrays: Investigating Latency and Energy Trade-offs versus Mesh and Torus Interconnects for Block-Level Matrix Multiplication Workloads
On a many-core chip, cores spend much effort passing data to each other. This work proposes a mirror-symmetric network-on-chip topology for 64-core arrays, with long-range bypass links, and compares it with 2D mesh and 2D torus networks in a simulator for block-level matrix multiplication workloads.
-
AI-PASS: An AI-Based Predictive, Adaptive and Energy Aware Task Scheduling Framework for Heterogeneous Parallel Computing
Modern computers mix CPUs, GPUs and specialised devices, and workloads keep changing. This paper introduces AI-PASS, a scheduling framework that predicts workload and adapts task and resource selection, including an Adaptive Task Suitability Score, for heterogeneous parallel computing.
-
Comparative Analysis of Classical, Metaheuristic, and Reinforcement Learning Approaches to Load Balancing in Cloud Computing
How should a cloud spread work across virtual machines of different power? This paper compares three families of load balancing, traditional heuristics, nature-inspired metaheuristics and reinforcement learning, on latency, throughput, energy, resource use and fault resilience, drawing on a review of peer-reviewed studies.
-
Next-Generation OS Architectures for Low-Power Smart Hardware: A Comprehensive Review
Standard operating systems like Linux do not fit very small, resource-limited devices. This review surveys operating systems for low-power smart hardware and presents a general framework for automatically creating lightweight, custom operating systems for application-specific hardware.
-
AI-Driven Malware Detection in Linux: A Survey of Techniques, Challenges, and Future Directions
Linux now runs cloud servers, IoT devices and embedded platforms, which makes it a target for malware. This survey reviews AI-based Linux malware detection research from 2022 to 2026, covering more than 40 methods from traditional machine learning to deep learning and transformer-based models.