SFS Turbo File System Types
This section describes the features, highlights, and application scenarios of different types of SFS Turbo file systems.
| Parameter | 20 MB/s/TiB | 40 MB/s/TiB | 125 MB/s/TiB | 250 MB/s/TiB | 500 MB/s/TiB | 1,000 MB/s/TiB |
| Max. bandwidth | Min. [20 × Storage capacity (TiB), 8,000] MB/s | Min. [40 × Storage capacity (TiB), 8,000] MB/s | Min. [125 × Storage capacity (TiB), 20,000] MB/s | Min. [250 × Storage capacity (TiB), 20,000] MB/s | Min. [500 × Storage capacity (TiB), 80,000] MB/s | Min. [1,000 × Storage capacity (TiB), 80,000] MB/s |
| Max. IOPS | 250,000 | 250,000 | 1 million | 1 million | 1 million | 1 million |
| Single-queue, 4 KiB latency | 2–5 ms | 2–5 ms | 1–3 ms | 1–3 ms | 1–3 ms | 1–3 ms |
| Capacity | 3.6 TB to 1 PB | 1.2 TB to 1 PB | 1.2 TB to 1 PB | 1.2 TB to 1 PB | 1.2 TB to 1 PB | 1.2 TB to 1 PB |
| Medium Type | HDD | HDD | SSD | SSD | ESSD | ESSD |
| Highlights | Large capacity and low cost | Large capacity and low cost | Low latency and cost effectiveness | Low latency and cost effectiveness | High IOPS and high-density performance | High IOPS and high-density performance |
| Typical use cases | Log storage, file sharing, content management, and websites | Log storage, file sharing, content management, and websites | AI training, autonomous driving, EDA simulation, rendering, enterprise NAS, and HPC web applications | AI training, autonomous driving, EDA simulation, rendering, enterprise NAS, and HPC web applications | Large-scale AI training, AI models, and AI generated content | Large-scale AI training, AI models, and AI generated content |
| Parameter | Standard | Standard-Enhanced (Discontinued) | Performance | Performance-Enhanced (Discontinued) |
| Max. bandwidth | 150 MB/s | 1 GB/s | 350 MB/s | 2 GB/s |
| Max. IOPS | 5,000 | 15,000 | 20,000 | 100,000 |
| Single-queue, 4 KiB latency | 2–5 ms | 2–5 ms | 1–3 ms | 1–3 ms |
| Capacity | 500 GB to 32 TB | 10 TB to 320 TB | 500 GB to 32 TB | 10 TB to 320 TB |
| Medium Type | HDD | HDD | SSD | SSD |
| Highlights | Large capacity and low cost | Low latency and high IOPS | ||
| Typical use cases | Code storage, log storage, file sharing, and enterprise OA | High-performance website, file sharing, content management, image rendering, AI training, and enterprise OA | ||
- For SFS Turbo Standard, Standard-Enhanced, Performance, and Performance-Enhanced file systems, performance is determined by the file system type and remains unaffected by the purchased capacity.
- For SFS Turbo 20 MB/s/TiB, 40 MB/s/TiB, 125 MB/s/TiB, 250 MB/s/TiB, 500 MB/s/TiB, and 1,000 MB/s/TiB file systems, performance scales with the purchased capacity. As the purchased capacity increases, the available bandwidth also increases. Based on the bandwidth provided by each 1 TiB of capacity, SFS Turbo file systems are classified into the following types: 20 MB/s/TiB, 40 MB/s/TiB, 125 MB/s/TiB, 250 MB/s/TiB, 500 MB/s/TiB, and 1,000 MB/s/TiB.
For example, if you buy a 250 MB/s/TiB file system with a capacity of 6 TiB, the bandwidth that this file system can provide is 1,500 MB/s (250 × 6).
The minimum bandwidth of an SFS Turbo file system is 150 MB/s. If the calculated bandwidth is less than 150 MB/s, 150 MB/s will be the bandwidth of this file system. There will also be a maximum bandwidth for each type of SFS Turbo file system. If the calculated bandwidth is greater than the maximum bandwidth, the maximum bandwidth will be the bandwidth of this file system.
- In the table, the maximum IOPS and maximum bandwidth all include both the read and write operations. So, maximum IOPS = read IOPS + write IOPS.
- The minimum expansion increment of an SFS Turbo Standard-Enhanced (discontinued), Standard, Performance-Enhanced (discontinued), or Performance file system is 100 GB. The minimum expansion increment of a 20 MB/s/TiB, 40 MB/s/TiB, 125 MB/s/TiB, 250 MB/s/TiB, 500 MB/s/TiB, or 1,000 MB/s/TiB file system is 1.2 TB, and the new capacity must be an integer multiple of 1.2 TB.
Feedback
Was this page helpful?
Provide feedbackThank you very much for your feedback. We will continue working to improve the documentation.