Blended Storage

In recent years, emerging storage hardware technologies have focused on divergent goals: better performance or lower cost-per-bit:

Correspondingly, data systems that employ these technologies are typically optimized either to be fast (but expensive) or cheap (but slow). We take a different approach: by architecting a storage system to natively utilize two tiers of fast and low-cost storage technologies, we can achieve a Pareto-efficient balance between performance and cost-per-bit.

​This project sets to fundamentally rethink data structures, and algorithms for caching, prefetching and migration for multi-tiered storage.

---

Publications

RubbleDB: CPU-Efficient Replication with NVMe-oF, Haoyu Li, Sheng Jiang, Chen Chen, Ashwini Raina, Xingyu Zhu, Changxu Luo, Asaf Cidon. ATC 2023.

Efficient Migrations Between Storage Tiers with PrismDB, Ashwini Raina, Jianan Lu, Asaf Cidon, Michael J. Freedman. ASPLOS 2023.

Artifacts

PrismDB source code.

Team

Faculty / senior researchers: Asaf Cidon, Michael J. Freedman
PhD students: Ashwini Raina, Jianan Lu, Haoyu Li

Teaching

Fall 2021: class project in EECS 6897, distributed storage systems

Funding