Intent-Driven Storage Systems: From Low-Level Tuning to High-Level Understanding

TL;DR

The paper addresses the challenge of optimizing storage systems with an immense configuration space by introducing Intent-Driven Storage Systems (IDSS), which leverages large language models (LLMs) to infer workload intent from unstructured signals and coordinate cross-layer configurations within safety guardrails. It defines four design principles—autonomous adaptation, holistic optimization, guarded autonomy, and vendor-neutral policy abstraction—and a four-phase architecture (Data Acquisition, Data Organization, Reasoning, Configuration) to translate telemetry into validated, platform-agnostic policies. Empirical results on FileBench show significant IOPS improvements (up to $2.45\times$) and real-trace performance approaching the best policy (geometric mean around 97%), demonstrating the feasibility of semantic, intent-driven optimization across storage components. Collectively, the work suggests a path toward adaptive, autonomous storage systems that align with dynamic workload goals by bridging administrators, applications, and hardware through a unified, intent-aware framework.

Abstract

Existing storage systems lack visibility into workload intent, limiting their ability to adapt to the semantics of modern, large-scale data-intensive applications. This disconnect leads to brittle heuristics and fragmented, siloed optimizations. To address these limitations, we propose Intent-Driven Storage Systems (IDSS), a vision for a new paradigm where large language models (LLMs) infer workload and system intent from unstructured signals to guide adaptive and cross-layer parameter reconfiguration. IDSS provides holistic reasoning for competing demands, synthesizing safe and efficient decisions within policy guardrails. We present four design principles for integrating LLMs into storage control loops and propose a corresponding system architecture. Initial results on FileBench workloads show that IDSS can improve IOPS by up to 2.45X by interpreting intent and generating actionable configurations for storage components such as caching and prefetching. These findings suggest that, when constrained by guardrails and embedded within structured workflows, LLMs can function as high-level semantic optimizers, bridging the gap between application goals and low-level system control. IDSS points toward a future in which storage systems are increasingly adaptive, autonomous, and aligned with dynamic workload demands.

Figures (4)

• Figure 1: Overview of the IDSS framework for system-wide parameter optimization.
• Figure 2: Synthetic traces. Hatched bar shows the policy selected by the LLM.
• Figure 3: Cache policy performance and distribution for real-world block I/O traces.
• Figure 4: Comparison of different methods of storage system parameter tuning on different workloads.