Adaptive data placement cuts ssd write overhead on shared servers
eBPF-Driven Adaptive Lifetime-Aware Data Placement for Shared FDP SSDs
Operating Systems
Summary
SSDs slow down and wear out faster when they repeatedly rewrite data in the same places. To help, Flexible Data Placement (FDP) separates data by how long it will be kept, reducing unnecessary work. The authors of this paper created eFDP, a system that automatically and continuously learns how long different data lives and adapts where data is stored on SSDs used by many tenants at once. eFDP improves performance and reduces extra writing, even as tenants change over time.
What this means in practice
- •For storage system engineers: Automatically optimize data storage layouts in multi-tenant SSDs to reduce maintenance overhead and improve writing speed.
- •For database administrators: Improve write performance and SSD lifespan of databases sharing storage by dynamically adjusting data placement based on data lifetime.
Authors
Hajin Moon, Taewan Noh, Chaeeun Yoon, Jeonghun Lee, Junyoung Han, Joo-Young Hwang, Youngjae Kim
Abstract
NVMe Flexible Data Placement (FDP) reduces SSD write amplification by separating data with different lifetimes to minimize garbage collection (GC) copying. However, on shared servers, multiple tenants generate more lifetime classes than the available placement handles. Existing approaches rely on static mappings or application-specific knowledge, failing to capture runtime lifetimes and adapt to changing workloads. The challenge is to identify which data die together and determine which classes should be separated to maximize GC savings under a limited handle budget. We present eFDP, an eBPF-based system that automatically learns data lifetimes and manages FDP placement for unmodified applications on a stock Linux kernel. By tracking writes and invalidations, eFDP identifies lifetime classes and estimates the GC copying saved by separating them. It first assigns each tenant a handle, then allocates the remaining handles to the separations offering the greatest savings. It also re-plans placement when active tenants change. With four database tenants on an emulated FDP SSD, eFDP eliminates twice as much excess write amplification as tenant-level isolation and improves every writing tenant's throughput by 19-23%. On an enterprise FDP SSD, given RocksDB's levels, it matches a manually optimized placement using three handles instead of five. After a tenant switch, eFDP re-plans within seven minutes and maintains 25% lower write amplification than the original layout even after an hour.