Liquid-State Drive: A Case for DNA Block Device for Enormous Data
summary
LIQUID-STATE DRIVE implements a DNA-backed block device using small fast SSDs plus a simulator prototype and claims concrete reductions: write amplification falls up to seven orders of magnitude (and 2,927× in traces), read cost drops up to 6,206× (and 7× in traces). The design uses a dual-layer translation table with L0 on SSD, stores GC metadata symbiotically with physical blocks to avoid DNA metadata writes, and applies delayed invalidation to remove costly DNA reads from the update critical path.
AI-written summary · about
chapters (6)
- 0:00DNA storage basics
- 1:46DNA vs traditional storage
- 3:33DNA block device integration
- 7:06Liquid State Drive design
- 10:40Performance evaluation
- 12:28Conclusion
related talks
18mThe Solid State Storage Revolution
45mBig Data and SSDs
30mAlgorithms for DNA Data Storage
14mRe-think Data Management Software Design Upon the Arrival of Storage Hardware with
16mZNS: Avoiding the Block Interface Tax for Flash-based SSDs
13mClosing the B+-tree vs. LSM-tree Write Amplification Gap on Modern Storage Hardware...
53mOvercome Blast Radius Anxiety With Ultrafast Rebuilds and Performance for High-capacity SSDs
06mNano-node: Intelligent Hard Drives in Large Storage Infrastructures