Hybrid Workflow Composition for Extreme-Scale Data Processing: A Case Study on the HL-LHC (Extended Version)

A Structure-Aware Hybrid Scheduling Framework for Mixed-Dependency Workflow Scheduling in V2X Testing

Efficient workflow scheduling is essential for improving testing efficiency in large-scale Vehicle-to-Everything (V2X) protocol conformance testing. Existing directed acyclic graph (DAG) schedulers treat all test components uniformly, overlooking the inherent structural heterogeneity of V2X workflows which contain mixed independent and precedence-constrained components. This leads to cross-subsystem interference, low resource utilization, and extended makespan. This paper presents AOE–CP (AON DAG with Edge-Weighted Transformation and Critical Path Scheduling), a structure-aware hybrid scheduling architecture for V2X testing. Unlike existing heuristic-based improvements, AOE–CP achieves performance gains through domain-specific structural reorganization rather than new scheduling rules. It integrates three mechanisms: (i) structural decomposition to decouple independent and dependent components and resolve their conflicting optimization objectives; (ii) critical path thread isolation to eliminate cross-subsystem resource competition and guarantee zero critical path waiting time for V2X dominant-topology workflows; (iii) atomic operation-based time estimation supporting fully offline scheduling without runtime profiling overhead. Experiments show that, for large-scale workflow scenarios, AOE–CP reduces normalized makespan by 14–16% and 22–24% compared with HEFT (Heterogeneous Earliest Finish Time) and CPOP (Critical Path On Processor), respectively. Scheduling overhead is only 2.1 ms for 1000-component workflows. AOE-CP reaches makespan saturation with four threads, versus 12 for HEFT and 28 for CPOP, demonstrating superior resource efficiency and scalability. The framework can also be generalized to other mixed-dependency workflow scenarios.

Intelligent Workflow Scheduling for Distributed Data Processing Systems

An intelligent workflow scheduling framework that improves performance through adaptive decision-making, predictive analytics, and machine learning, and addresses key challenges like load balancing, scalability, energy efficiency, and fault tolerance is proposed.

Enhancing the Kubernetes Scheduler: A State-of-the-Art Review from Cloud to Edge

A comprehensive review of Kubernetes scheduling strategies published between January 2023 and January 2026 is presented and a multi-dimensional taxonomy is established that categorizes scheduling approaches based on common objectives, modification methods, optimization methodologies, targeted workloads, evaluation methods, scheduling scopes, and performance metrics.

A Unified Bandwidth Orchestration Framework for Hierarchical Data Storage Systems

G-STAR: Graph-based Scheduling with Trace-driven Adaptive Routing for Industrial LLM-based Multi-Agent Systems

Large Language Model-based Multi-Agent Systems (LLM-MAS) have shown exceptional promise for complex tasks, including retrieval-augmented generation and autonomous data analytics. However, their deployment in resource-constrained industrial environments faces critical challenges, such as unpredictable end-to-end latency, low task success rates, and expensive operational costs. Existing orchestration strategies for LLM-MAS mainly rely on static heuristics, implicit LLM-based routing, or reinforcement learning, which suffer from brittleness under workload drift, prohibitive online exploration costs, and inherent governance risks. To address these issues, we propose G-STAR, a general graph-based scheduling framework that formalizes complex MAS pipelines as attributed Directed Acyclic Graphs (DAGs). Specifically, G-STAR decouples its learning pipeline into three critical stages to ensure both system stability and execution efficiency. First, we build a data layer that logs fine-grained signals for agent node (e.g., execution latency, token I/O, model IDs, resource metrics) and task-grounded outcomes (e.g., accuracy, cost, SLA), yielding a trace-driven dataset. Second, we train an offline GNN model to predict optimal dynamic routing decisions, including agent activation, concurrency-constrained Top-K selection, and execution ordering, conditioned on the evolving graph state. Finally, a lightweight dispatcher deploys the pre-trained GNN model for single-forward-pass routing with negligible computational overhead. Extensive experiments on two public benchmarks and our real-world production workload confirm G-STAR's superiority over static and LLM-driven baselines. It maintains competitive task quality while reducing latency variability and consistently improving end-to-end latency. Furthermore, we develop an industry-grade orchestration stack with asynchronous execution, resilient serving, and audit-friendly artifacts, offering a practical solution for optimizing web-scale deployments of complex MAS pipelines.

Workflow-Aware Expert Routing for Distributed LLM Serving Over the Edge-Cloud Continuum

Deploying Large Language Models (LLMs) over the edge-cloud continuum faces severe stability challenges due to the conflict between stochastic network topology and complex workflow dependencies. Existing schedulers, relying either on computationally prohibitive Graph Neural Networks (GNNs) or topology-agnostic heuristics, fail to reconcile this tension. To bridge these gaps, we propose STEM, a service-level and topology-aware orchestration framework that formulates distributed LLM serving as a workflow-aware routing problem over a monitored service overlay, in which heterogeneous service instances act as specialized experts. At the core of STEM lies the STAR-PPO algorithm, utilizing a lightweight graph-free perception mechanism. By leveraging Squeeze-and-Excitation attention, it extracts critical bottleneck features from raw telemetry with linear complexity, bypassing the scalability limits of message-passing paradigms. To further achieve Pareto-efficient trade-offs, we develop a Dynamic Weight Adaptation (DWA) mechanism that autonomously recalibrates optimization preferences based on entropy-regularized metric drift. Extensive experiments on real-world datasets spanning 2,000 nodes demonstrate that our framework significantly outperforms state-of-the-art baselines. Specifically, STAR-PPO reduces network transmission costs by 96.8% and improves comprehensive inference efficiency by 24.4%, while sustaining robust zero-shot generalization across regions, with average latency within $1.09\times $ of a target-domain-retrained reference under a strict cross-region protocol. Code and data are available at https://github.com/gymorsiback/STARPPO