Similar articles

Abstract: Emergence of new hardware, including persistent memory and smart network interface card (SmartNIC), has brought new opportunities to file system design. In this paper, we design and implement a new file system named NICFS based on persistent memory and SmartNIC. We divide the file system into two parts: the front end and the back end. In the front end, data writes are appended to the persistent memory in a log-structured way, leveraging the fast persistence advantage of persistent memory. In the back end, the data in logs are fetched, processed, and patched to files in the background, leveraging the processing capacity of SmartNIC. Evaluation results show that NICFS outperforms Ext4 by about 21%/10% and about 19%/50% on large and small reads/writes, respectively.

NICFS：基于持久化内存和智能网卡的文件系统

[1]Anderson TE, Canini M, Kim J, et al., 2019. Assise: performance and availability via NVM colocation in a distributed file system. https://arxiv.org/abs/1910.05106

[2]Chen YM, Lu YY, Zhu BH, et al., 2021. Scalable persistent memory file system with kernel-userspace collaboration. Proc 19^th USENIX Conf on File and Storage Technologies, p.81-95.

[3]Condit J, Nightingale EB, Frost C, et al., 2009. Better I/O through byte-addressable, persistent memory. Proc ACM SIGOPS 22^nd Symp on Operating Systems Principles, p.133-146.

[4]Gray C, Cheriton D, 1989. Leases: an efficient fault-tolerant mechanism for distributed file cache consistency. ACM SIGOPS Oper Syst Rev, 23(5):202-210.

[5]Kim J, Jang I, Reda W, et al., 2021. LineFS: efficient SmartNIC offload of a distributed file system with pipeline parallelism. Proc ACM SIGOPS 28^th Symp on Operating Systems Principles, p.756-771.

[6]Lee C, Sim D, Hwang JY, et al., 2015. F2FS: a new file system for flash storage. Proc 13^th USENIX Conf on File and Storage Technologies, p.273-286.

[7]Li JR, Lu YY, Wang Q, et al., 2022. AlNiCo: SmartNIC-accelerated contention-aware request scheduling for transaction processing. Proc USENIX Annual Technical Conf, p.951-966.

[8]Liang Z, Lombardi J, Chaarawi M, et al., 2020. DAOS: a scale-out high performance storage stack for storage class memory. Proc 6^th Asian Conf on Supercomputing Frontiers, p.40-54.

[9]Lu YY, Shu JW, Zheng WM, 2013. Extending the lifetime of flash-based storage through reducing write amplification from file systems. Proc 11^th USENIX Conf on File and Storage Technologies, p.257-270.

[10]Lu YY, Shu JW, Wang W, 2014. ReconFS: a reconstructable file system on flash storage. Proc 12^th USENIX Conf on File and Storage Technologies, p.75-88.

[11]Lu YY, Shu JW, Chen YM, et al., 2017. Octopus: an RDMA-enabled distributed persistent memory file system. Proc USENIX Annual Technical Conf, p.773-785.

[12]Lu YY, Shu JW, Zhang JC, 2019. Mitigating synchronous I/O overhead in file systems on open-channel SSDs. ACM Trans Stor, 15(3):17.

[13]NVIDIA, 2022. DOCA. https://developer.nvidia.com/networking/doca [Accessed on Oct. 8, 2022].

[14]Ou JX, Shu JW, Lu YY, 2016. A high performance file system for non-volatile main memory. Proc 11^th European Conf on Computer Systems, Article 12.

[15]Schuh HN, Liang WH, Liu M, et al., 2021. Xenic: SmartNIC-accelerated distributed transactions. Proc ACM SIGOPS 28^th Symp on Operating Systems Principles, p.740-755.

[16]Xu J, Swanson S, 2016. NOVA: a log-structured file system for hybrid volatile/non-volatile main memories. Proc 14^th USENIX Conf on File and Storage Technologies, p.323-338.

[17]Zhang JC, Shu JW, Lu YY, 2016. ParaFS: a log-structured file system to exploit the internal parallelism of flash devices. Proc USENIX Annual Technical Conf, p.87-100.