RSRL — Rutgers Systems Research Lab

We build scalable, efficient, and secure systems for handling the rising heterogeneity of modern memory and storage hardware.

Department of Computer Science, Rutgers University · CoRE 313 · PI: Sudarsun Kannan

Team

Faculty

Sudarsun Kannan

Associate Professor

Operating systems, memory & storage

PhD Students

David Domingo

PhD Student

Storage reliability · VM scheduling

Jian Zhang

PhD Student

Direct-access FS · near-storage acceleration

Linfeng He

PhD Student

Redesigning virtual memory

Haoyu Li

PhD Student

Devon Lewis

PhD Student

Long Tran

PhD Student

Edge micro-datacenters

Minrui Tian

PhD Student

Linjie Ma

PhD Student

Disaggregated storage caching

Abhilash Nambissan

PhD Student

Undergraduates

Sumanta Das

Undergraduate

Alumni

Yujie Ren — PhD · File system scalability, virtual memory · now in industry
Krishnan Gosakan — Visiting Researcher · Fast address translation for virtual memory
Shaleen Garg — PhD (with Manish Parashar) · HPC systems
Sreeram Maddineni — MS · Distributed heterogeneity · Amazon
Jae Woo Joo — MS · Amazon
Avin Abraham — BS · Memory heterogeneity
Kyle Stratton — BS
Benjamin Watkins — BS · Graph problems
Rohit Rao — BS · Systems for ML
Aayush Agnihotri — High School · Memory management

Research Areas

File Systems Memory Heterogeneity Operating Systems Storage Virtual Memory Persistent Memory Direct-Access I/O Near-Storage Machine Learning Caching Prefetching TLB Paging Datacenter CXL Heterogeneous Storage Wildfire Monitoring Hazard Detection Edge Smart Storage Disaggregation Address Translation KV-Store LSM Durability Concurrency Multi-tenancy Kernel Reliability HPC

Direct-access and Smart Storage

Bypassing OS layers and offloading file-system functionality to firmware and near-storage accelerators while preserving consistency, integrity, and security.

FAST '18–'25, OSDI '20, SOSP '21

OmniCache FusionFS CrossFS PolyStore

Memory Management and Heterogeneity

Operating systems and runtimes for heterogeneous memory (DRAM, NVM, CXL, stacked 3D), including page management, address translation, and TLB scaling.

ISCA '17 – ASPLOS '23, SPAA '21–'25, TALG '25

Mosaic Pages HeteroOS KLOCs

I/O Prefetching and Caching

Cross-layered prefetching, context-aware near-storage prefetching, and collaborative client–server caches for disaggregated storage.

ASPLOS '24, HotStorage '24–'25

CrossPrefetch Context-aware Prefetching

File System Correctness and Reliability

Parallel checkers and repair tools, comprehensive testing for configuration-rich KV stores, and userspace NVM file systems with security guarantees.

FAST '21–'23, SOSP '23, SYSTOR '25

pFSCK Trio (Best Paper) PANGOLIN

Data Durability and Transactions

Persistent transactional memory, durable concurrency, and fault-tolerant NVM data structures.

ASPLOS '20, FAST '23

TimeStone TENET

Edge and Geo-Distributed Heterogeneity

Edge-centric micro-datacenters, multitenancy at the edge, and resource heterogeneity in distributed deployments.

HotStorage '20–'25, HotInfra '24, MILCOM '25

Edge MicroDCs

Selected Publications

Full publication list → · Google Scholar · DBLP

Horizontally Scaling Heterogeneous Storage for Combined Capabilities with PolyStore

Yujie Ren, David Domingo, Jian Zhang, Rekha Pitchumani, Sanidhya Kashyap, Sudarsun Kannan

ACM Transactions on Storage (ToS), 2026

Don't Melt Your Cache: Low-Associativity with Heat-Sink

Michael A. Bender, Alex Conway, Daniel DeLayo, Martin Farach-Colton, Jaehyun Han, Linfeng He, Rob Johnson, Sudarsun Kannan, William Kuszmaul, Donald E. Porter, Evan West

ACM SPAA 2025 Distinguished Paper Award

Paper

Kamino: Efficient VM Allocation at Scale with Latency-Driven Cache-Aware Scheduling

David Domingo, Hugo Barbalho, Marco Molinaro, Kuan Liu, Abhisek Pan, David Dion, Thomas Moscibroda, Sudarsun Kannan, Ishai Menache

USENIX OSDI 2025

Paper Slides

PolyStore: Exploiting Combined Capabilities of Heterogeneous Storage

Yujie Ren, David Domingo, Jian Zhang, Paul John, Rekha Pitchumani, Sanidhya Kashyap, Sudarsun Kannan

USENIX FAST 2025

Paper Code

OmniCache: Collaborative Caching for Near-storage Accelerators

Jian Zhang, Yujie Ren, Marie Nguyen, Changwoo Min, Sudarsun Kannan

USENIX FAST 2024

Paper Code

Enabling High-Performance and Secure Userspace NVM File Systems with the Trio Architecture

Diyu Zhou, Vojtech Aschenbrenner, Tao Lyu, Jian Zhang, Sudarsun Kannan, Sanidhya Kashyap

ACM SOSP 2023 Best Paper Award

Mosaic Pages: Big TLB Reach with Small Pages

Krishnan Gosakan, Jaehyun Han, William Kuszmaul, et al., Sudarsun Kannan, Donald Porter

ACM ASPLOS 2023 Distinguished Paper Award

Paper Code

Software

Open-source research artifacts from our group. Browse all repos at the RutgersCSSystems GitHub organization.

File & Storage Systems

PolyStore — heterogeneous storage with combined capabilities (FAST '25)
FusionFS — firmware file system fusing I/O ops via CISCOps (FAST '22)
CrossFS — cross-layered direct-access file system (OSDI '20)
pFSCK — parallel filesystem checking and repair (FAST '21)
Trio / ArckFS — userspace NVM file system (SOSP '23, '25)

Near-Storage Acceleration

OmniCache — collaborative caching for near-storage accelerators (FAST '24)
CrossPrefetch — cross-layered I/O prefetching (ASPLOS '24)

Memory & Address Translation

KLOCs — kernel-level object contexts for heterogeneous memory (ASPLOS '21)
Mosaic Pages (gem5) — big TLB reach with small pages (ASPLOS '23)
NoveLSM — LSM key-value store for NVM (ATC '18)
TENET — fault-tolerant persistent transactional memory (FAST '23)
TimeStone — durable transactional memory (ASPLOS '20)

Edge & Wildfire Monitoring

HazardMon — stateful edge triage for wildfire monitoring (MILCOM '25)
Fire-YOLOv5 — wildfire detection model for edge devices

Research Support

Our work is supported by the following organizations:

Contact

Sudarsun Kannan — sudarsun.kannan [at] cs.rutgers.edu
Office: CoRE 313, Department of Computer Science, Rutgers University
110 Frelinghuysen Rd, Piscataway, NJ 08854

PI homepage · GitHub (RutgersCSSystems) · Google Scholar