fault-tolerant design

This study addresses the growing challenge of runtime failures in locally deployed and fine-tuned open-source large language models (LLMs), which increasingly stem not from algorithmic flaws but from systemic fragility in the deployment stack. Through a large-scale empirical analysis of 705 real-world failure reports from the DeepSeek, Llama, and Qwen ecosystems, we identify and formally characterize three recurring phenomena: diagnostic divergence, system homogeneity, and lifecycle upgrade issues. By integrating fault log analysis, root cause categorization, and cross-ecosystem comparison, we construct the first reliability analysis framework for open-source LLM deployment, establishing the deployment stack as the primary source of failures. We further release the first public dataset of such failures, offering actionable insights to enhance deployment reliability.

This work addresses the reliability degradation and silent failures commonly observed in tool-augmented large language models, which often stem from orchestration-layer issues such as tool timeouts, incorrect parameters, or stale context. The authors formulate reliability as a bounded runtime control problem and introduce the first self-healing orchestration framework that integrates fault classification, budget-constrained action selection, and verification-guided recovery. By leveraging runtime observability to detect failure signals, classify fault types, and execute targeted recovery strategies within resource budgets, the system achieves a 98.8% single-attempt recovery success rate on a 100-task benchmark—outperforming naive retry and full replanning by 8.7 and 5.8 percentage points, respectively. Notably, the approach entirely eliminates semantic silent failures, substantially enhancing both system reliability and diagnosability.

This work addresses the trade-off between reliability and cost in large language model (LLM) agents for tool invocation: fully LLM-driven routing incurs high computational overhead, while static workflows are prone to silent failures under compound tool errors. To overcome this, we propose the Self-Healing Router architecture, which introduces runtime fault tolerance into graph-based tool-calling systems for the first time. By modeling control flow as a weighted graph routing problem, our approach employs parallel health monitoring to dynamically assess tool status and leverages Dijkstra’s algorithm for deterministic shortest-path rerouting, invoking the LLM only when necessary. Evaluated across 19 scenarios, the method achieves accuracy comparable to ReAct while reducing control-plane LLM calls by 93% (from 123 to 9) and completely eliminating silent failures, thereby enabling binary observability and efficient autonomous recovery.

This paper addresses the challenges of high configuration complexity and excessive human dependency in end-to-end Time-Sensitive Networking (TSN) orchestration, which hinder its adoption in real-time safety-critical systems. To this end, we propose the first Large Language Model (LLM)-assisted automated TSN orchestration framework. Our method integrates an LLM with a Network Calculus (NC) engine, network simulation tools, and external verification modules to enable closed-loop generation—from natural-language requirements to formally verifiable TSN configurations. Key contributions include: (1) pioneering the application of LLMs to deterministic networking orchestration; (2) designing an extensible, TSN-specific architectural blueprint and implementation roadmap; (3) empirically validating the feasibility of current LLMs for TSN configuration tasks; and (4) identifying critical future directions—including TSN-specific datasets and benchmark suites—thereby bridging a fundamental gap in LLM applications to deterministic networks.

High-performance software systems accumulate latent reliability risks through aggressive optimizations; superficial performance metrics (e.g., high cache hit rates) mask underlying bottlenecks, leading to load amplification and cascading failures upon degradation. Current reliability engineering emphasizes reactive mitigation, lacking proactive identification and prevention of optimization-induced fragility. Method: We propose the first systematic framework for optimization-risk management, introducing a novel quantitative model and the Latent Risk Index (LRI). Our tripartite defense architecture—HYDRA (risk detection), RAVEN (perturbation-based validation), and APEX (risk-aware optimization)—integrates mathematical modeling, six categories of optimization-sensitive perturbation testing, and high-precision online monitoring. Contribution/Results: Experiments demonstrate 89.7% risk detection rate, >92.9% monitoring accuracy, 69.1% reduction in MTTR, annual cost savings of $1.44M, and a payback period of just 3.2 months.

This study addresses the challenges of high latency, unstable concurrency, and security risks faced by large language model (LLM) agents in automating asset lifecycle management within Industry 4.0. The authors propose a Plan-then-Execute architecture that generates verifiable workflow graphs and integrates a topology-aware parallel scheduling mechanism to enable controlled inference overlap while ensuring functional correctness and security. Key technical contributions include topological-sort-based multi-agent scheduling, structured context pruning, dependency-aware concurrency control, and graceful degradation under fault injection. Evaluated on the AssetOpsBench benchmark, the system reduces median end-to-end latency by 1.6× (up to 1.8× for highly parallel tasks) and cuts inference overhead by approximately 30% through context pruning, all while maintaining stable task completion rates and output quality.

This study systematically investigates the capability boundaries of large language models (LLMs) in security tool orchestration, with a focus on the relative impact of model choice, client implementation, toolset composition, and reasoning mechanisms on system performance. Leveraging the open-source orchestration framework HexStrike-AI, the authors conduct multi-configuration comparative experiments across 86 picoCTF challenges, complemented by failure diagnosis and targeted refinements—including tool corrections, behavioral adjustments, and capability extensions—to quantitatively demonstrate, for the first time, the critical role of the client component in determining the performance of a fixed LLM. Results indicate that performance bottlenecks primarily stem from reasoning or environmental constraints rather than missing tools, enabling an increase in overall solve rate from 55.4% to 72.0% (p < 0.001) with high reproducibility (17 out of 20 trials consistent). The work introduces a reproducible evaluate-and-improve feedback loop, establishing a new paradigm for intelligent security agent systems.

This study addresses the challenge of automating workflows in complex industries—such as logistics, healthcare, and construction—where processes are fragmented across heterogeneous tools and involve multi-party collaboration. The work proposes orchestration as a core abstraction to enable effective automation by dynamically coordinating multi-step tasks, enforcing domain-specific constraints, managing human approvals, and integrating legacy systems. It introduces the novel concept of “orchestration bottlenecks” and develops a theoretical framework that unifies multi-agent systems, workflow modeling, constraint reasoning, and human–AI collaboration, while exposing critical gaps in current multi-agent approaches at the orchestration level. Based on distinct sources of operational friction across domains, the paper advocates for targeted architectural safeguards—such as constraint enforcement or explainability—and phased implementation strategies to provide actionable pathways for automation in complex operational environments.

While microservice failures are readily detectable, root cause analysis remains inefficient due to alarm flooding and the absence of structured memory capturing system dependencies and historical behaviors. This work proposes a topology-aware, operation-memory-driven multi-agent architecture that decouples root cause inference from explanation for the first time: the former relies on deterministic computation using a learned dependency graph and temporal anomaly thresholds, while the latter leverages a large language model to generate interpretable recommendations grounded in structured evidence. A novel four-layer operational memory mechanism enables traceable and reusable autonomous operations. Evaluated on an e-commerce benchmark platform with eight types of injected faults, the approach successfully reproduces and resolves two real-world cascading failures, significantly improving diagnostic accuracy and efficiency.

This work addresses the fragmentation in existing frameworks that treat deterministic and probabilistic computations in isolation, lacking a unified declarative language to orchestrate large language models (LLMs) and symbolic tools. We propose Structured Prompt Language (SPL), the first framework to deeply integrate probabilistic operations (GENERATE/EVALUATE) and deterministic reasoning (SOLVE/ASSERT) within a single declarative paradigm. SPL supports shared variable binding, runtime dynamic routing, and seamless interoperability with LLMs (e.g., Ollama, Anthropic), symbolic engines (e.g., SymPy, SageMath, Lean), and the distributed execution grid Momagrid. Across 1,200 experiments, SPL achieves machine-verified correctness rates of 82–93% (e.g., 93% for gemma4:e2b), substantially outperforming pure LLM baselines; most failures stem from solver kernels rejecting invalid expressions.