Since its inception 12 years ago, serverless computing (specifically, FaaS) has transitioned from a radical trend to an infrastructure staple. The vast inefficiencies inherent in the serverless model gave rise to a vast body of research. Alas, many promising ideas turned out to be overly disruptive and failed to see industry adoption,
leading to a view that the serverless research landscape has plateaued.
In this talk, I will argue that the rigidity of commercial FaaS infrastructure is not a dead end for research, but rather an opportunity to thrive within confines with exciting use cases, system challenges, and research opportunities. I will demonstrate how we can build impactful systems not by trying to rewire the provider’s fabric, but by designing intelligent application-layer solutions.
Boris Grot is a Professor in the School of Informatics at the University of Edinburgh, where he leads the Edinburgh Architecture and Systems (EASE) Lab. Boris’s research seeks to bridge the gap between hardware and software systems with a focus on next-generation datacenters and the edge. A number of his labs’ outputs have been adopted by the industry. Boris is a member of the MICRO and HPCA Halls of Fame and a recipient of multiple awards for his research. Boris was the Program Chair for MICRO 2022, General Chair for HPCA 2024, and currently sits on the Board of Directors of ACM SIGARCH. He has a PhD from The University of Texas at Austin.
Danny Harnik is a Senior Technical Staff Member (STSM) at IBM Research, specializing in storage systems, data compression, cloud infrastructure and AI inference platforms. Over the years, he has played a key role in advancing IBM products across storage, cloud platforms, and open-source technologies. Danny has authored more than 35 academic publications and holds over 45 patents. He earned his PhD in cryptography from the Weizmann Institute of Science in Israel.
Sessions:
Day 1 Zoom-TBD or Live – systor 2026-TBD
Opening Session
Chair: Anand Sivasubramaniam (Pennsylvania State University)
Session 1: OS, Concurrency & Observability
Chair: Javier Gonzalez (Samsung)
- NAGC: Node-Aware Garbage Collection for F2FS
Xiaole He (Shandong University of Technology); Xin Cui (Shandong University of Technology); Lihua Yang (National University of Defense and Technology); Xuyang Wang (Shandong University of Technology); Sisi Zhang (Shandong University of Technology); Qian Zhang (Shandong University of Technology); Xinchang Bi (Shandong University of Technology) - Boundary Conditions for Semantic Log Parsing in Systems Observability Pipelines
Dumo Ngwenya (The Open University UK); Patrick Wong (The Open University UK); Dhouha Kbaier (The Open University UK) - Mitigating False Collisions in Partition-Based Concurrency Control
Matthew Curtis-Maury (NetApp, Inc.); Sushrut Bhowmik (NetApp, Inc.); Grey Files (NetApp, Inc.); Jian Hu (NetApp, Inc.); Bryan Prosser (NetApp, Inc.)
Session 2: LLM Serving
Chair: Avishay Traeger (Red Hat)
- Astrolabe: Balancing Load in LLM Serving with Randomized Prediction-Guided Scheduling
Wei Da (University of Cambridge); Evangelia Kalyvianaki (University of Cambridge) - To Keep or Not to Keep: Learning KV Cache Retention in Disaggregated LLM Serving Systems
Dong Liu (University of California, LA); Yanxuan Yu (Columbia University); Eric Jiang (University of California, LA); Shu Wang (University of California, LA); Ying Nian Wu (University of California, LA)
Day 2 Zoom-TBD or Live – systor 2026-TBD
Session 3: Intelligent Storage Systems
Chair: Assaf Natanzon (Eon.io)
- BASIC-Prefetcher: Bin-based Address and Size-Informed Caching for AI-Driven SSD Workloads
Han Jang (Oregon State University); Dongjun Lee (Oregon State University); Youngbin Jin (Solidigm, Inc); Sean Bullis (Oregon State University); Ben Lee (Oregon State University) - TellyTune: Exploring the Limits of LLM-Assisted Tuning for LSM-Based Key-Value Stores
Viraj Thakkar (Arizona State University); Qi Lin (Arizona State University); Kenanya Prasetyo (ITS); Raden Haryosatyo Wisjnunandono (ITB); Jin Wang (Arizona State University); Achmad Imam Kistijantoro (ITB); Reza Fuad Rachmadi (ITS); Zhichao Cao (Arizona State University) - Mario: Hierarchically Pipelined Approximate Nearest Neighbor Search on Computational Storage Devices
Kaijiang Deng (Shanghai Jiao Tong University); Shengan Zheng (Shanghai Jiao Tong University); Penghao Sun (Shanghai Jiao Tong University); Guifeng Wang (Shanghai Jiao Tong University); Weihan Kong (Shanghai Jiao Tong University); Linpeng Huang (Shanghai Jiao Tong University)
Session 4: Storage Architectures
Chair: Rohan Mehta (Micron)
- When Do Zoned Namespaces SSDs Matter? A Comparative Study of Cache Workloads
John Ramsden (University of British Columbia); Sam Cheng (University of British Columbia); Hans Holmberg (Western Digital); Matias Bjørling (Western Digital); Alexandra Fedorova (University of British Columbia and MongoDB) - 4Z-SSD: A Semantic-Aware Four-Zone SSD Architecture for Balancing Write Amplification and Flash-Level Parallelism
Dongjun Lee (Oregon State University); Han Jang (Oregon State University); Youngbin Jin (Solidigm, Inc); Ben Lee (Oregon State University) - Revisiting Block-interface ZNS RAID Systems with Dynamically-Sized RAID (DS-RAID) and Random Accessible Unified Memory (RAUM) Xiangqun Zhang (Syracuse University); Bryan S. Kim (Syracuse University)
Poster Session:
Chair: Sarel Cohen (Reichman University)
For accepted posters, see
