Background EC is mainly for cold storage, providing low redundancy costs and availability. While hot storage need to keep QoS, so expensive replication is more common.
ECWide uses wide stripe to decrease redundancy while avoid repairing bandwidth waste.
(SIGMOD ‘18) HYMEM: Managing Non-Volatile Memory in Database Systems
How to predict future hotness of a new request in photo cache systems?
RocksDB based on LSM-Tree is widely used in Facebook. This work aims to model and analysis RocksDB workloads at Facebook.
This work points out that how to utilize low-level performance of a hierarchical storage system.
This work focuses on the bottleneck of file-mapping in NVM filesystems.
Intro With the development of storage device, lots of novel stuff come to world, including NAND SSD, Optane 3D XPoint, STT-RAM… Personal machines have already multi-level storage: SRAM on CPU L2/L2 cache, DRAM on CPU’s DIMMs, SSD to accelerate operating system, and HDD to storage big file like media, movie.
Paper notes about Assise, a distributed FS on NVM. Assise keeps data & metadata in clients (opposite to disaggregation) to enble fast recovery and high perf.
paper notes about XStore from SJTU, which uses learned cache (aka learned index) for distributed DRAM KVS.