• Publications
  • Influence
NVMKV: A Scalable and Lightweight Flash Aware Key-Value Store
NVMKV is a scalable and lightweight KV store that leverages advanced capabilities that are becoming available in modern FTLs and outperforms state of the art KV stores by 50%-300%, and significantly improves performance predictability for the YCSB KV benchmark when compared with the popular LevelDB KV Store. Expand
NVMKV: A Scalable, Lightweight, FTL-aware Key-Value Store
NVMKV is a lightweight KV store that leverages native FTL capabilities such as sparse addressing, dynamic mapping, transactional persistence, and support for high-levels of lock free parallelism. Expand
Write policies for host-side flash caches
Two consistent write-back caching policies, ordered and journaled, are developed and evaluated that are designed to perform increasingly better than write-through and can trade-off staleness for performance, approaching, and in some cases, exceeding conventional write- back performance. Expand
ANViL: Advanced Virtualization for Modern Non-Volatile Memory Devices
A new form of storage virtualization based on block-level address remapping that enables a variety of useful features and optimizations to be readily implemented, including snapshots, deduplication, and single-write journaling is presented. Expand
ECO: Harmonizing Edge and Cloud with ML/DL Orchestration
This work presents Edge Cloud Orchestrator (ECO), an architecture for enabling realistic ML deployments that leverage both edge and cloud by providing an abstraction to orchestrate, manage, and automate ML pipelines. Expand
An Analysis of Error Behaviour in a Large Storage System
This paper analyzes the error behavior of a 3.2TB disk storage system. We report reliability data for 18 months of the prototype''s operation, and analyze 6 months of error logs from nodes in theExpand
Snapshots in a flash with ioSnap
IoSnap represents a case study of how to integrate snapshots into a modern, well-engineered flash-based storage system and incurs negligible performance overhead during normal operation. Expand
The Art of Massive Storage: A Web Image Archive
The Thinker ImageBase, the San Francisco Fine Arts Museums' online art image database, demonstrates issues involved in managing large storage systems and delivering their contents to users. Expand
HEC: improving endurance of high performance flash-based cache devices
This paper analyzes the added write pressures that cache workloads place on flash devices and proposes optimizations at both the cache and flash management layers to improve endurance while maintaining or increasing cache hit rate. Expand
Don't Stack Your Log On My Log
It is shown that multiple log layers affects sequentiality and increases write pressure to flash devices through randomization of workloads, unaligned segment sizes, and uncoordinated multi-log garbage collection. Expand