A Diffusion-Based Framework for Configurable and Realistic Multi-Storage Trace Generation

2025-09-01

AI Summary

Existing storage trace generation methods suffer from low fidelity, poor configurability, and insufficient modeling of cross-device temporal dependencies. To address these limitations, this paper proposes DiTTO—the first diffusion-based framework for multi-device storage trace generation. DiTTO integrates time-series modeling with conditional guidance to support fine-grained user specifications (e.g., I/O access patterns, device load distributions) and explicitly captures inter-device temporal dependencies. Experimental evaluation demonstrates that DiTTO significantly outperforms state-of-the-art approaches: it achieves high fidelity (only 8% error), enhanced diversity (+32% Jensen–Shannon distance), and strong configuration consistency (>96% requirement matching rate). By generating highly realistic, customizable, and temporally coherent synthetic traces, DiTTO establishes a trustworthy foundation for storage system simulation, testing, and optimization.

Technology Category

• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Search and Optimization: Sampling/Simulation-based Search
• Data Mining & Knowledge Management: Representing, Reasoning, and Using Provenance, Trust

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Web performance, measurement, and characterization
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Web Mining and Content Analysis: Web data generation and simulation

Abstract

We propose DiTTO, a novel diffusion-based framework for generating realistic, precisely configurable, and diverse multi-device storage traces. Leveraging advanced diffusion tech- niques, DiTTO enables the synthesis of high-fidelity continuous traces that capture temporal dynamics and inter-device dependencies with user-defined configurations. Our experimental results demonstrate that DiTTO can generate traces with high fidelity and diversity while aligning closely with guided configurations with only 8% errors.

Problem

• Generates configurable multi-device storage traces
• Captures temporal dynamics and inter-device dependencies
• Ensures high fidelity with user-defined configurations

Innovation

• Diffusion-based framework for multi-storage traces
• Generates configurable traces with temporal dynamics
• Achieves high fidelity with 8% configuration error