The underlying hardware for flash memory-based solid-state drive (SSD) is becoming increasingly unreliable, and the current state-of-the-art approaches implement performance-sacrificing techniques to prevent data loss. This problem is exacerbated as the storage ages, causing performance unpredictability and making performance anomaly detection difficult. In the current capacity-invariant interface where the storage device exports a fixed capacity throughout its lifetime, an SSD has no other choice but to trade performance for reliability as it wears out. This project rethinks the existing storage interface from a holistic perspective by exploiting tradeoffs among capacity, performance, and reliability (CPR) and designing a flexible capacity-variant interface that allows the SSD to maintain performance while it gracefully reduces the number of exported blocks.
The project involves three research thrusts. The first thrust is to quantify the error-induced performance degradation by building an SSD aging framework. The second thrust then builds a capacity-variant system and demonstrates the effectiveness of such an approach. Lastly, the project studies how capacity-variance can be extended across a heterogeneous aggregation of SSDs. The project is expected to have a major impact in the storage industry, as the problem it addresses is both timely and likely to get worse if left unchecked. A capacity-variant storage system can make storage system maintenance and provisioning easier, and this project will help educate students on the value and importance of cross-layer optimization and holistic design thinking.
This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
