FAST '20
as the organisers put it
18th USENIX Conference on File and Storage Technologies (FAST 20), February 24–27, 2020, Hyatt Regency Santa Clara, Santa Clara, CA, USA. FAST brings together storage-system researchers and practitioners to explore new directions in the design, implementation, evaluation, and deployment of storage systems.
who it's for
storage-system researchers, storage engineers, cloud infrastructure engineers, database engineers, site reliability engineers, systems performance engineers, firmware and device engineers
topics
CRUSH-based object placement extensions; operation inference for cloud sync; FPGA computational storage drives; storage parameter importance analysis; flash-optimized read-as-needed search; copy-on-abundant-write cloning; Lustre I/O workload characterization; coupon-based I/O fairness mechanism; manycore SSD firmware design; enterprise SSD reliability study; erasure-coded Raft variants; RocksDB workload benchmarking
| title | author | length |
|---|---|---|
| MAPX: Controlled Data Migration in the Expansion of Decentralized Object Based Storage | Li Wang, Yiming Zhang, Jiawei Xu, Guangtao Xue | 17m |
MAPX extends CRUSH by adding a virtual layer beneath root for each expansion and controlling Placement Group placement via static and dynamic PG timestamps… | ||
| Lock-Free Collaboration Support for Cloud Storage Services with Operation Inference | Jian Chen, Minghao Zhao, Zhenhua Li, Ennan Zhai, Feng Qian, Hongyi Chen, Yunhao Liu, Tianyin Xu | 16m |
Reverse engineering and measurements reveal long synchronization tails, missing locks, and many false conflicts in popular cloud drives. The system infers… | ||
| POLARDB Meets Computational Storage: Efficiently Support Analytical Workloads in Cloud | Wei Cao, Yang Liu, Zhushi Cheng, Ning Zheng, Wei Li, Wenjie Wu, Linqiang Ouyang, Peng Wang, Yijing Wang, Ray Kuan, Zhenjun Liu, Feng Zhu, Tong Zhang | 38m |
Paper reports a deployed replacement of SSDs with FPGA-based computational storage drive devices inside Alibaba Cloud POLARDB, plus coordinated changes to the… | ||
| Carver: Finding Important Parameters for Storage System Tuning | Zhen Cao, Geoff Kuenning, Erez Zadok | 28m |
Carver measures each storage parameter’s effect as the variance-based metric (reduction in response variance), samples huge spaces with Latin Hypercube… | ||
| Read as Needed: Building WiSER, a Flash-Optimized Search Engine | Jun He, Kan Wu, Sudarsun Kannan, Andrea C. Arpaci-Dusseau, Remzi H. Arpaci-Dusseau | 31m |
WiSER is a clean-slate, 11k-line C++ search engine that read as needed from high-performance SSDs using four concrete techniques: grouping by term layout, a… | ||
| How to Copy Files | Yang Zhan, Alex Conway, Yizheng Jiao, Nirjhar Mukherjee, Ian Groombridge, Michael A. Bender, Martin Farach-Colton, William Jannen, Rob Johnson, Donald E. Porter, Jun Yuan | 30m |
A concrete implementation in BetrFS 0.5 converts Bε-trees into a Bε-DAG and delivers logical clones by delaying physical copies until many writes accumulate, a… | ||
| Uncovering Access, Reuse, and Sharing Characteristics of I/O-Intensive Files | Tirthak Patel, Suren Byna, Glenn K. Lockwood, Nicholas J. Wright, Philip Carns, Robert Ross, Devesh Tiwari | 26m |
Analysis of Cori production logs (84M Darshan records, 52M files, ~13 PB total) uses Darshan and Lustre metadata to isolate I/O‑intensive files (>=100 GB and… | ||
| GIFT: A Coupon Based Throttle-and-Reward Mechanism for Fair and Efficient I/O Bandwidth | Tirthak Patel, Rohan Garg, Devesh Tiwari | 34m |
GIFT implements a coupon-based throttle-and-reward system that throttles some jobs and later redeems coupons to restore fairness, using an LP optimizer that… | ||
| Scalable Parallel Flash Firmware for Many-core Architectures | Jie Zhang, Miryeong Kwon, Michael Swift, Myoungsoo Jung | 30m |
DeepFlash claims a concrete result: a manycore SSD firmware that reaches 1M IOPS and about 4.5 GB/s by splitting firmware into modular “stresses” and running a… | ||
| A Study of SSD Reliability in Large Scale Enterprise Storage Deployments | Stathis Maneas, Kaveh Mahdaviani, Tim Emami, Bianca Schroeder | 28m |
Field study analyzes 1.4 million SSDs from NetApp filers over 2.5 years (3 makers, 18 models, 12 capacities, SLC/cMLC/eMLC/3D‑TLC) and reports annual… | ||
| Making Disk Failure Predictions SMARTer! | Sidi Lu, Bing Luo, Tirthak Patel, Yongtao Yao, Devesh Tiwari, Weisong Shi | 27m |
Claims 0.95 F-measure and 0.95 MCC for a 10-day horizon, trained on 380,000 drives across 64 sites with SMART, disk-level performance and four-level location… | ||
| DC-Store: Eliminating Noisy Neighbor Containers using Deterministic I/O Performance | Miryeong Kwon, Donghyun Gouk, Changrim Lee, Byounggeun Kim, Jooyoung Hwang, Myoungsoo Jung | 29m |
DC-Store claims deterministic I/O and removes noisy-neighbor effects by physically dividing an SSD into four NVM sets (prototype uses 2 channels per set on a 2… | ||
| GoSeed: Generating an Optimal Seeding Plan for Deduplicated Storage | Aviv Nachman, Gala Yadgar, Sarai Sheinvald | 25m |
GoSeed formulates seeding as an ILP formulation and ships three practical accelerations: solver timeout, fingerprint sampling, and container aggregation. On… | ||
| Characterizing, Modeling, and Benchmarking RocksDB Key-Value Workloads at Facebook | Zhichao Cao, Siying Dong, Sagar Vemuri, David H.C. Du | 29m |
Open-sourced trace collector captures RocksDB GET/PUT/DELETE/MERGE/ITERATOR traces from three Facebook workloads (UDB, ZippyDB, UP2X) and analyzers reveal that… | ||
| FPGA-Accelerated Compactions for LSM based Key-Value Store | Teng Zhang, Jianying Wang, Xuntao Cheng, Hao Xu, Nanlong Yu, Gui Huang, Tieying Zhang, dengcheng he, Feifei Li, Wei Cao, Zhongdong Huang, Jianling Sun | 27m |
Offloads LSM compactions to FPGA co‑processors integrated with X-Engine, implementing pipelined decoder–merger–encoder compaction units over PCIe and running… | ||
| HotRing: A Hotspot-Aware In-Memory Key-Value Store | Jiqiang Chen, Liang Chen, Sheng Wang, Guoyun Zhu, Yuanyuan Sun, Huan Liu, Feifei Li, Le Cai | 24m |
HotRing claims a 2.58× throughput win on highly skewed workloads by reordering hash chains into an ordered-ring hash and advancing per-bucket head pointers… | ||
| An Empirical Guide to the Behavior and Use of Scalable Persistent Memory | Jian Yang, Juno Kim, Morteza Hoseinzadeh, Joseph Izraelevitz, Steven Swanson | 29m |
Large-scale experiments (10,000 microbenchmarks, ~300 machine-hours on Intel-supplied 2-socket hardware with 3 TB Optane made from 256 GB DIMMs) show Optane… | ||
| BCW: Buffer-Controlled Writes to HDDs for SSD-HDD Hybrid Storage Server | Shucheng Wang, Ziyi Lu, Qiang Cao, Hong Jiang, Jie Yao, Yuanyuan Dong, Puyuan Yang | 29m |
Claims HDDs can serve many writes at μs-level by exploiting periodic buffered-write phases; experiments on five production HDDs show 35–180μs fast-phase writes… | ||
| InfiniCache: Exploiting Ephemeral Serverless Functions to Build a Cost-Effective Memory | Ao Wang, Jingyuan Zhang, Xiaolong Ma, Ali Anwar, Lukas Rupprecht, Dimitrios Skourtis, Vasily Tarasov, Feng Yan, Yue Cheng | 31m |
InfiniCache runs entirely on AWS Lambda, shards objects with Reed‑Solomon (k+m) erasure coding, and coordinates via a proxy and client library to stream chunks… | ||
| Quiver: An Informed Storage Cache for Deep Learning | Abhishek Vijaya Kumar, Muthian Sivathanu | 30m |
Quiver is a cache layer co‑designed with PyTorch that uses secure hash-based addressing to let different jobs and accounts reuse identical files, implements… | ||
| CRaft: An Erasure-coding-supported Version of Raft for Reducing Storage Cost | Zizhong Wang, Tongliang Li, Haixia Wang, Airan Shao, Yunren Bai, Shangming Cai, Zihan Xu, Dongsheng Wang, Xiaoqi Chen | 27m |
CRaft replaces full-entry Raft replication with Reed-Solomon erasure coding (typical setting k=3, n=5 or 7) and builds a CRaft-backed key-value store. It… | ||
| Strong and Efficient Consistency with Consistency-Aware Durability | Aishwarya Ganesan, Ramnatthan Alagappan, Andrea C. Arpaci-Dusseau, Remzi H. Arpaci-Dusseau | 33m |
Introduces consistency-aware durability (CAD) and implements it by changing ZooKeeper into ORCA modifies ZooKeeper to enforce durability on reads rather than… | ||
| Hybrid Data Reliability for Emerging Key Value Storage Devices | Rekha Pitchumani, Yang-suk Kee | 29m |
KVMD implements a hybrid reliability manager evaluated on prototype Samsung NVMe KV SSDs and supports 4–255 byte keys and up to 2 MB values. It plugs hashing… | ||
| nothing matches | ||