SSDFS: ZNS SSD ready file system with zero GC overhead
summary
Claims a ZNS-aware SSDFS that achieves zero GC overhead by organizing storage into segments with per-segment mapping tables and bitmaped compaction, implementing extent-style logical-to-physical mapping and online migration of stale blocks. A prototype implements trim handling, continuous collector migration and update-safe superblock writes; tested against ZNS drives with a 100 GB workload and shows removed background GC and lower write amplification, extending SSD lifetime while keeping runtime updates possible.
AI-written summary · about
chapters (9)
- 0:00Introduction to SSDFS
- 1:13SSDFS design and structure
- 3:48Logical blocks and segments
- 6:50Mapping and metadata tables
- 9:07Migration and updates
- 12:05Migration conflicts and logic
- 15:31Trim requests and garbage collection
- 17:15Invalidation and data updates
- 20:02System overview and conclusion
related talks
14mZNS+: Advanced Zoned Namespace Interface for Supporting In-Storage Zone Compaction
20mZNSwap: un-Block your Swap
31mFrom Open-Channel SSDs to Zoned Namespaces
15mD2FS: Device-Driven Filesystem Garbage Collection
14mUnleashing Zoned UFS: Cross-Layer Optimizations for Next-Generation Mobile Storage
21mThe Design and Implementation of a Capacity-Variant Storage System
29mDC-Store: Eliminating Noisy Neighbor Containers using Deterministic I/O Performance
16mRearchitecting Buffered I/O in the Era of High-Bandwidth SSDs