Sponsored By



In Cooperation With

Supporters

Gold Sponsors

SK hynix

Silver Sponsors

Google

DELL Technologies

Follow us



Logo design by Michal O. Zadok.



Full Program

Day 1: Monday, July 8th, 2024

13:00 - 13:15 PDT Opening Remarks
13:15 - 14:30 PDT Session 1: Data Encoding and Compression
14:30 - 14:50 PDT Coffee Break
14:50 - 15:40 PDT Session 2: LLMs to the rescue
15:40 - 16:30 PDT Panel: Storage and Machine Learning
16:30 - 17:20 PDT Session 3: Considering Workloads

Day 2: Tuesday, July 9th, 2024

09:00 - 10:00 PDT Keynote by Carlos Maltzahn
10:00 - 11:15 PDT Session 4: Keeping Data Correct
11:15 - 11:30 PDT Coffee Break
11:30 - 12:20 PDT Session 5: Minimizing Latencies
12:20 - 14:00 PDT Lunch
14:00 - 15:15 PDT Session 6: Storage and Machine Learning
15:15 - 16:30 PDT Session 7: Storage Technologies
16:30 - 16:45 PDT Concluding Thoughts



Keynote Talks

Cross - Lessons learned in building up a center for research in open source sw
Carlos Maltzahn
Abstract:

While cloud computing is transforming society, today's public clouds are black boxes, implemented and operated by a single provider that makes all business and technology decisions. In 2014 we launched the Mass Open Cloud (MOC) with the vision of creating a production cloud that would enable innovation by a broad industry and research community. This open cloud has become a laboratory for cloud research and innovation, resulting in 100s of publications, contributions to open source software, and collaborations between researchers, open source developers, and production operations staff. This talk will discuss some of the storage research and initiatives that have gone on in the MOC ecosystem, our future directions, and some of the demands we see on storage going forward.

Bio:

Orran Krieger is a Professor in the Department of Electrical and Computer Engineering at Boston University. Despite an ongoing effort to focus on research and avoid responsibility, he is the Co-Director of the Red Hat Collaboratory@BU and is involved in a set of interrelated Open Cloud projects, including the Mass Open Cloud (MOC), Open Research Cloud Initiative (ORCI), and Open Cloud Testbed (OCT). Before coming to BU as a Fellow of the Hariri Institute for Computing, he spent five years at VMware starting and working on vCloud. Prior to that, he was a researcher and manager at IBM T. J. Watson, leading the Advanced Operating System Research Department, working on the K42 operating system and contributing to projects like PHYP, Cell, PERCS, rHype, Linux, sHype and even contributing a couple of instructions to the Power architecture. Orran obtained his PhD and MASc degrees in Electrical Engineering at the University of Toronto where he worked on the Hurricane and Tornado Operating Systems and the Hector and NUMAchine multiprocessors.



Panels

Storage and Machine Learning - What can we learn from each other?

Sessions

Session 1: Data Encoding and Compression

Is Low Similarity Threshold A Bad Idea in Delta Compression?
Hongming Huang, Chun Jason Xue, Nan Guan, Hong Xu

Dictionary Based Cache Line Compression
Daniel Cohen, Sarel Cohen, Dalit Naor, Daniel Waddington, Moshik Hershcovitch

Revisiting Erasure Codes: A Configuration Perspective
Runzhou Han, Chao Shi, Tabassum Mahmud, Zeren Yang, Vladislav Esaulov, Lipeng Wan, Yong Chen, Jim Wayda, Matthew Wolf, Mai Zheng


Session 2: LLMs to the rescue

Can Modern LLMs Tune and Configure LSM-based Key-Value Stores?
Viraj Thakkar, Madhumitha Sukumar, Jiaxin Dai, Kaushiki Singh, Zhichao Cao

ION: Navigating HPC I/O Optimization Journey using Large Language Models
Chris Egersdoerfer, Arnav Sareen, Jean Luca Bez, Suren Byna, Dong Dai


Session 3: Considering Workloads

Context-aware Prefetching for Near-Storage Accelerators
Jian Zhang, Marie Nguyen, Sanidhya Kashyap, Sudarsun Kannan

Advocating for Key-Value Stores with Workload Pattern Aware Dynamic Compaction
Heejin Yoon, Jin Yang, Juyoung Bang, Sam H. Noh, Young-ri Choi


Session 4: Keeping Data Correct

Secure Archival is Hard... Really Hard
Christopher Smith, Maliha Tabassum, Soumya Chowdary Daruru, Gaurav Kulhare, Arvin Wang, Ethan L. Miller, Erez Zadok

Asymmetric RAID: Rethinking RAID for SSD Heterogeneity
Ziyang Jiao, Bryan S. Kim

Shadow Filesystems: Recovering from Filesystem Runtime Errors via Robust Alternative Execution
Jing Liu, Xiangpeng Hao, Andrea Arpaci-Dusseau, Remzi Arpaci-Dusseau, Tej Chajed


Session 5: Minimizing Latencies

Breaking Barriers: Expanding GPU Memory with Sub-Two Digit Nanosecond Latency CXL Controller
Donghyun Gouk, Seungkwan Kang, Hanyeoreum Bae, Eojin Ryu, Sangwon Lee, Dongpyung Kim, Junhyeok Jang, Myoungsoo Jung

Improving Virtualized I/O Performance by Expanding the Polled I/O Path of Linux
Dongjoo Seo, Yongsoo Joo, Nikil Dutt


Session 6: Storage and Machine Learning

Rethinking Erasure-Coding Libraries in the Age of Optimized Machine Learning
Jiyu Hu, Jack Kosaian, K. V. Rashmi

A Selective Preprocessing Offloading Framework for Reducing Data Traffic in DL Training
Meng Wang, Gus Waldspurger, Swaminathan Sundararaman

Quantitative Analysis of Storage Requirement for Autonomous Vehicles
Yuxin Wang, Yuankai He, Ruijun Wang, Weisong Shi


Session 7: Storage Technologies

Can ZNS SSDs be Better Storage Devices for Persistent Cache?
Chongzhuo Yang, Zhang Cao, Chang Guo, Ming Zhao, Zhichao Cao

Can Storage Devices be Power Adaptive?
Dedong Xie, Theano Stavrinos, Kan Zhu, Simon Peter, Baris Kasikci, Thomas E. Anderson

Life-after-Death: Exploring Thermal Annealing Conditions to Enhance 3D NAND SSD Endurance
Matchima Buddhanoy, Sudeep Pasricha, Biswajit Ray