Lune

FAST2023Top-tier venue

Fisc: A Large-scale Cloud-native-oriented File System

Qiang Li, Lulu Chen, Xiaoliang Wang, Shuo Huang, Qiao Xiang, Yuanyuan Dong, Wenhui Yao, Minfei Huang, Puyuan Yang, Shanyang Liu, Zhaosheng Zhu, Huayong Wang

2023Year
18Citations
7Top-tier citations

Abstract

Despite the progress of cloud-native technologies, existing distributed file systems are ill-suited for multi-tenant cloudnative applications for two reasons, First, their clients are typically heavyweight, resulting in a low level of resource multiplexing among containers. Second, their architecture is based on network gateway and falls short in providing efficient, highly-available and scalable I/O services for cloud-native applications. In this paper, we propose Fisc, a large-scale, cloudnative-oriented distributed file system. Fisc introduces three key designs: (1) a lightweight file system client to improve the multiplexing of resources with a two-layer resource aggregation, (2) a storage-aware distributed gateway to improve the performance, availability and scalability of I/O services, and (3) a DPU-based virtio-Fisc device offloading key functions to hardware. Fisc has been deployed in production for over three years and now serves cloud applications running over 3 million cores. Results show that Fisc client only consumes 69% CPU resources compared to the traditional file system client. The production environment shows that the average latency of online searching tasks is less than 500 µs when they access their files in Fisc.

Ask about this paper

Your agent reads all of it.

Lune indexed this paper to the last equation, along with the top-tier papers that cite it. Ask a question and the answer quotes them.

Questions to start from

Your agent calls

Luneget_paper_fulltext

Ask in Lune

Free to start. No credit card required.

lune papers fulltext 1cde17e4-7e99-46ef-ad9c-dadd8f4742e8

Cited by top-tier papers7

Ask how each one uses it

Builds on19

Related papers

Dusk over the sea between two cliffs drawn in fine vertical lines