SHF: Small: Reengineering Database Systems for Fast SSDs

University Of California-San Diego, La Jolla CA

Investigators

Abstract

A new class of non-volatile, solid-state memories (e.g., phase-change memory, spin-torque MRAMs, and the memristor) are emerging that promise to revolutionize the way that computer systems store and access data. While these memory offer significant promise they will also require us to re-engineer many parts of existing computing systems. No where is this more critical than in database systems, where overall throughput (in terms of transactions per seconds) is closely tied to the performance of the underlying storage system and the optimizations that have been applied in software. This project is analyzing the implications for these new memories on database systems and devising now hardware and software mechanisms to improve performance, reduce energy consumption, and improve reliability. The potential impacts of these optimizations is wide-reaching. Database (in various forms) constitute the heart of the cloud computing infrastructure that supports many of the "killer apps" that are driving technologies forward. Leveraging these new technologies, will make it cheaper, easier, and greener to implement existing applications and will enable new applications that are not currently possible.