SplitFUSE: a highly compatible framework for I/O acceleration of user-level file systems

Abstract

User-level file systems are widely adopted in research and production environments due to their flexibility and reduced risk of kernel crashes. Filesystem in Userspace (FUSE) is a general-purpose framework for developing user-level file systems in Linux. Compared with library-based file systems, FUSE adopts a cooperative architecture between kernel modules and userspace libraries, ensuring metadata security and compliance with standard portable operating system interface (POSIX) semantics. However, this architecture introduces substantial context-switching overhead. Existing optimization approaches for high-performance computing environments often improve input/output (I/O) performance at the expense of FUSE's cross-environment compatibility and kernel-level security guarantees. To address this limitation, this study proposes SplitFUSE, an I/O-acceleration framework for user-level file systems that preserves high compatibility and strong security. SplitFUSE introduces a split architecture that decouples metadata and data-request processing. Specifically, the kernel maintains full metadata consistency, while the userspace retains only a minimal metadata subset required to validate and process data requests that bypass the kernel securely. Implemented as a self-contained mechanism, SplitFUSE preserves the same cross-environment compatibility as conventional FUSE. Experimental results demonstrate that, under I/O-intensive small-write workloads, SplitFUSE achieves up to 4-6 times higher write bandwidth than native FUSE and outperforms state-of-the-art alternatives. For common file system workloads, it delivers substantial performance improvements with minimal migration overhead.

Keywords

Filesystem in Userspace (FUSE) / User-level file system / Input/Output (I/O) acceleration / Kernel bypass / Split architecture

References

[1] An W , Bi X , Chen GT , et al., 2024. Fire-flyer AI-HPC:a cost-effective software-hardware co-design for deep learning. Proc Int Conf for High Performance Computing, Networking, Storage and Analysis, Article 83.

[2] Bijlani A , Ramachandran U , 2019. Extension framework for file systems in user space. Proc USENIX Annual Technical Conf, p.121- 134.

[3] Cho KJ , Choi J , Kwon H , et al., 2024. RFUSE:modernizing userspace filesystem framework through scalable kernel-userspace communication. Proc 22^nd USENIX Conf on File and Storage Technologies, Article 9.

[4] Cornell B , Dinda P , Bustamante F , 2004. Wayback:a user-level versioning file system for Linux. Proc USENIX Annual Technical Conf, p.19- 28.

[5] Deng TL , Chen C , Yin SY , 2022. On the efficiency of user-level parallel file systems. J Univ Chin Acad Sci, 39 (2): 275- 282.

[6] Huai QB , Hsu W , Lu JW , et al., 2021. XFUSE:an infrastructure for running filesystem services in user space. Proc USENIX Annual Technical Conf, p.863- 875.

[7] Jia WQ , Jiang DJ , Xiong J , 2025. A high-performance and scalable userspace log-structured file system for modern SSDs. ACM Trans Storage, 21 (4): 30.

[8] Lawrence Livermore National Laboratory , 2015. ZFS.https://zfsonlinux.org[Accessed on Mar. 2, 2026].

[9] Lembke J , Roman PL , Eugster P , 2022. DEFUSE:an interface for fast and correct user space file system access. ACM Trans Storage, 18 (3): 22.

[10] Mashtizadeh A , Bittau A , Huang Y , et al., 2013. Replication, history, and grafting in the Ori file system. Proc 24^th Symp on Operating Systems Principles, p.151- 166.

[11] Pontes R , Burihabwa D , Maia F , et al., 2017. SafeFS:a modular architecture for secure user-space file systems:one FUSE to rule them all. Proc 10^th ACM Int Systems and Storage Conf, Article 9.

[12] Red Hat , 2019. Gluster. https://gluster.org[Accessed on Mar. 2, 2026].

[13] Ren K , Gibson G , 2013. TABLEFS:enhancing metadata efficiency in the local file system. Proc USENIX Conf on Annual Technical Conf. p.145- 156.

[14] Sigurbjarnarson H , Ragnarsson P , Yang JC , et al., 2016. Enabling space elasticity in storage systems. Proc 9^th ACM Int Systems and Storage Conf, Article 6.

[15] Trapexit , 2014. Mergerfs.https://github.com/trapexit/mergerfs[Accessed on Mar. 2, 2026].

[16] Ungureanu C , Atkin B , Aranya A , et al., 2010. HydraFS:a high-throughput file system for the HYDRAstor content-addressable storage system. Proc 8^th USENIX Conf on File and Storage Technologies. p.225- 239.

[17] Weil S , Brandt S , Miller E , et al., 2006. Ceph:a scalable, high-performance distributed file system. Proc 7^th USENIX Symp on Operating Systems Design and Implementation. p.307- 320.

[18] Yan WR , Yao J , Cao Q , 2019. DeFUSE:decoupling metadata and data processing in FUSE framework for performance improvement. IEEE Access, 7: 138473- 138484.

[19] Zhang J , Ren Y , Kannan S , 2022. FusionFS:fusing I/O operations using CISCOps in firmware file systems. Proc 20^th USENIX Conf on File and Storage Technologies. p.297- 312.

[20] Zhu Y , Wang T , Mohror K , et al., 2018. Direct-FUSE:removing the middleman for high-performance FUSE file system support. Proc 8^th Int Workshop on Runtime and Operating Systems for Supercomputers, Article 6.

Rights & permissions

The Authors. Published by Zhejiang University Press Co., Ltd.