Flexible Data Placement on SSDs for Database Systems

Traditionally solid-state drives (SSDs) does not give the users the ability to control the data placement on the SSD. This often leads to suboptimal performance and lowers SSD lifetime, since SSDs internally don’t allow in-place updates. The updated disk pages are written elsewhere and the old versions have to be garbage collected. This poses problems if data with different lifetimes and update frequencies often gets written in the same region.

To cope with this issue, there has been several attempts by storage vendors, one of the most recent being flexible data placement (FDP) SSDs [1]. This new class of SSDs allow hints from programs to the SSD to dictate which data should be written where, thereby allowing separation of data from different applications or with different lifetimes.

A recent MSc thesis [2] at ITU integrated FDP support in DuckDB and tested in on real FDP SSDs. A follow-up thesis evaluated the performance of it further. However, the availability of real SSDs with FDP support is limited. This project aims at analyzing the FDP performance on DuckDB using a recently developed emulator [3] to overcome the availability challenge.

This project would be suitable as a standalone project, a research project with an MSc thesis followup, or BSc thesis. If you are interested in data management systems, storage devices, benchmarking, and performance analysis in general, this project would be a great fit for you. Furthermore, there is a possibility of collaborating with Samsung Research Denmark.

[1] Allison et al. “Towards Efficient Flash Caches with Emerging NVMe Flexible Data Placement SSDs.”, EuroSys 2025.

[2] Houlborg et al. “Flexible I/O for Database Management Systems with xNVMe.”, CIDR 2026.

[3] Song et al. “Characterizing and Emulating FDP SSDs with WARP.”, FAST 2026.