BUSE: Block Device in Userspace
| Název práce v češtině: | BUSE: Blokové zařízení v uživatelském prostoru |
|---|---|
| Název v anglickém jazyce: | BUSE: Block Device in Userspace |
| Klíčová slova: | linux|zařízení|ovladač |
| Klíčová slova anglicky: | linux|device|driver |
| Akademický rok vypsání: | 2021/2022 |
| Typ práce: | rigorózní práce |
| Jazyk práce: | angličtina |
| Ústav: | Katedra softwarového inženýrství (32-KSI) |
| Vedoucí / školitel: | Peter Desnoyers |
| Řešitel: | skrytý - zadáno a potvrzeno stud. odd. |
| Datum přihlášení: | 25.05.2022 |
| Datum zadání: | 25.05.2022 |
| Datum potvrzení stud. oddělením: | 25.05.2022 |
| Datum a čas obhajoby: | 19.04.2023 14:00 |
| Datum odevzdání elektronické podoby: | 25.05.2022 |
| Datum odevzdání tištěné podoby: | 25.05.2022 |
| Datum proběhlé obhajoby: | 19.04.2023 |
| Zásady pro vypracování |
| FUSE (File system in USEr space) is a widely-used kernel module and library allowing implementation of filesystems in userspace and is a very mature technology in Linux. This lets storage researchers rapidly prototype novel filesystems ideas without dealing with complexity of kernel code. FUSE-based file systems are even used in certain non-performance-critical production roles. User-space implementations of block devices, however, is poorly supported and much less common. Two technologies are used for this purpose, Network Block Device (NBD) and Target Control Mode in Userspace (TCMU). Neither scale with the number of CPUs in multi-core systems, and often exhibit unstable behaviour during configuration.
The goal of this thesis is to design, implement and experimentally evaluate a new technology for block device in userspace implementation (BUSE) on Linux. A key goal of the work will be achieving not only high single-core performance, but scalable performance on multi-core machines. Performance and scalability will be evaluated and compared with that of NBD and TCMU. An additional part of the thesis will be a demonstration of BUSE using network-remote S3 object storage, which will be compared to CloudBD, the current state-of-the-art solution. |
| Seznam odborné literatury |
| [1] Bjørling, Matias, et al. "Linux block IO: introducing multi-queue SSD access on multi-core systems." Proceedings of the 6th international systems and storage conference. 2013.
[2] Mickens, James, et al. "Blizzard: Fast, cloud-scale block storage for cloud-oblivious applications." 11th {USENIX} Symposium on Networked Systems Design and Implementation ({NSDI} 14). 2014. [3] Pan, Satadru, et al. "Facebook's Tectonic Filesystem: Efficiency from Exascale." 19th {USENIX} Conference on File and Storage Technologies ({FAST} 21). 2021. [4] Li, Huiba, et al. "Ursa: Hybrid block storage for cloud-scale virtual disks." Proceedings of the Fourteenth EuroSys Conference 2019. 2019. [5] Network Block Device, https://nbd.sourceforge.io [6] TCMU, https://www.kernel.org/doc/Documentation/target/tcmu-design.txt [7] Amazon Elastic Block Store (EBS), https://aws.amazon.com/ebs [8] CloudBD, https://www.cloudbd.io |