DDS: DPU-optimized Disaggregated Storage [Extended Report]
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866912004348837888 |
|---|---|
| author | Zhang, Qizhen Bernstein, Philip Chandramouli, Badrish Hu, Jiasheng Zheng, Yiming |
| author_facet | Zhang, Qizhen Bernstein, Philip Chandramouli, Badrish Hu, Jiasheng Zheng, Yiming |
| contents | This extended report presents DDS, a novel disaggregated storage architecture enabled by emerging networking hardware, namely DPUs (Data Processing Units). DPUs can optimize the latency and CPU consumption of disaggregated storage servers. However, utilizing DPUs for DBMSs requires careful design of the network and storage paths and the interface exposed to the DBMS. To fully benefit from DPUs, DDS heavily uses DMA, zero-copy, and userspace I/O to minimize overhead when improving throughput. It also introduces an offload engine that eliminates host CPUs by executing client requests directly on the DPU. Adopting DDS' API requires minimal DBMS modification. Our experimental study and production system integration show promising results -- DDS achieves higher disaggregated storage throughput with an order of magnitude lower latency, and saves up to tens of CPU cores per storage server. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2407_13618 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | DDS: DPU-optimized Disaggregated Storage [Extended Report] Zhang, Qizhen Bernstein, Philip Chandramouli, Badrish Hu, Jiasheng Zheng, Yiming Distributed, Parallel, and Cluster Computing H.2.4; D.4.2 This extended report presents DDS, a novel disaggregated storage architecture enabled by emerging networking hardware, namely DPUs (Data Processing Units). DPUs can optimize the latency and CPU consumption of disaggregated storage servers. However, utilizing DPUs for DBMSs requires careful design of the network and storage paths and the interface exposed to the DBMS. To fully benefit from DPUs, DDS heavily uses DMA, zero-copy, and userspace I/O to minimize overhead when improving throughput. It also introduces an offload engine that eliminates host CPUs by executing client requests directly on the DPU. Adopting DDS' API requires minimal DBMS modification. Our experimental study and production system integration show promising results -- DDS achieves higher disaggregated storage throughput with an order of magnitude lower latency, and saves up to tens of CPU cores per storage server. |
| title | DDS: DPU-optimized Disaggregated Storage [Extended Report] |
| topic | Distributed, Parallel, and Cluster Computing H.2.4; D.4.2 |
| url | https://arxiv.org/abs/2407.13618 |