TraceSynth: Generating Production-Quality Kernel Traces with Constraint-Guided Diffusion Models

2026-07-13

AI Summary

This work addresses the challenges of collecting real kernel execution traces in industrial systems, which are hindered by high overhead, storage costs, and privacy constraints. To overcome these limitations, we propose TraceSynth—the first framework to apply constraint-guided diffusion models to synthetic kernel trace generation. By leveraging multi-channel sequence modeling (capturing event types, timestamps, CPU affinity, etc.) and a lightweight repair mechanism, TraceSynth produces high-fidelity synthetic traces while preserving system semantic consistency. Experimental results demonstrate that on deterministic workloads such as scimark2, our approach achieves an F1-Macro score of 87.2% at a context length of L=4096, only 2.6 percentage points below that of real data. Constraint-based repair improves performance by up to 4.3%, and a two-channel lightweight variant retains 97–99% of full-model performance at half the computational cost.

Abstract

Machine learning models for system diagnostics rely on kernel execution traces to capture fine-grained system behavior, but collecting production traces in industrial systems is costly due to runtime overhead, storage demands, and privacy constraints. We present TraceSynth, a diffusion-based framework for generating synthetic kernel traces that augment limited real data for downstream ML tasks. TraceSynth models traces as multi-channel sequences (event types, timestamps, CPU affinity, thread identifiers, and process metadata) using a Transformer-based denoising diffusion process with constraint-guided repair to enforce system invariants. Across six benchmarks, results show strong workload dependence. For deterministic, compute-heavy workloads (scimark2), synthetic augmentation achieves 87.2% F1-Macro at context length L=4096, only 2.6 percentage points below real-only baselines. Context length is the dominant quality factor, with L=4096 yielding a +104% relative improvement over L=256, while constraint-guided repair improves synthetic data quality by up to 4.3%. Ablation studies show that lightweight 2-channel models retain 97-99% of the performance of full 6-channel models at roughly half the computational cost. TraceSynth supports cost-effective augmentation of kernel execution traces in production observability pipelines and helps identify when synthetic data can substitute for limited real traces.

Problem

• kernel traces
• production systems
• data collection cost
• privacy constraints
• system diagnostics

Innovation

• diffusion models
• kernel traces
• constraint-guided synthesis
• Transformer-based denoising
• synthetic data augmentation