This topic describes the performance metrics of File Storage NAS (NAS), the relationship between metrics, and the performance of each storage class.
Performance metrics
The performance of NAS file systems is measured by using the following metrics:
Input/output operations per second (IOPS)
IOPS measures the number of read and write operations that can be performed per second. The metric indicates the random read and write capabilities of NAS. If you need to perform a large number of random read and write operations, pay attention to the IOPS.
Throughput
Throughput measures the amount of data that can be transferred per second. Unit: MB/s. If you need to perform a large number of read and write operations, pay attention to the throughput.
Latency
Latency measures the amount of time that is required for a NAS file system to process an I/O request. Unit: milliseconds or microseconds. High latency may cause performance degradation or application errors. If your applications are sensitive to latency, we recommend that you use Extreme NAS file systems.
Relationship between IOPS, throughput, read and write block size, and latency
IOPS, throughput, read and write block size, and latency are tightly correlated. You can optimize the performance of a NAS file system to a certain extent by adjusting the size of read and write blocks and the number of concurrent tasks. For example, you can achieve a specific throughput and reduce the overall task latency by increasing the size of read and write blocks or increasing the number of concurrent tasks. At the same time, increasing the size of read and write blocks can prevent the IOPS from reaching the upper limit.
The following formulas show the relationship between the read and write block size, IOPS, throughput, number of reads and writes, and latency:
Throughput = IOPS × Read and write block size
IOPS = Number of reads and writes/Overall task latency
Overall task latency = Write latency × (Number of writes/Concurrency)
For example, you have a Capacity NAS file system whose write latency is about 100 ms per 1 MiB, 15 ms per 8 KiB, and 10 ms per 4 KiB, and the file system supports a maximum of 128 concurrent tasks. If you need to write 1 MiB of data per second to the file system, you can use the solutions described in the following table.
No. | Read and write block size | Concurrency | Number of writes | Overall task latency | IOPS | Throughput | Description |
Solution 1 | 4 KiB | 1 | 250 | 10 ms × 250 = 2.5s | 250/2.5s = 100 | 4 KiB × 100 = 400 KiB/s | Small read and write block size and low concurrency result in poor throughput and high latency. In this case, the throughput cannot reach 1 MiB/s. |
Solution 2 | 1 MiB | 1 | 1 | 100 ms | 1/0.1s = 10 | 1 MiB × 10 = 10 MiB/s | Compared with Solution 1, the read and write block size of Solution 2 is increased, and the throughput and latency performance are improved. The throughput can reach 1 MiB/s. However, the overall task latency is high. |
Solution 3 | 4 KiB | 125 | 250 | 10 ms × (250/125) = 20 ms | 250/0.02s = 12,500 | 4 KiB/s × 12,500 ≈ 49 MiB/s | Compared with Solution 1, the concurrency of Solution 3 is increased, and the throughput and latency performance are improved. The throughput can reach 1 MiB/s. The overall task latency is low, but the IOPS reaches the upper limit of the file system. |
Solution 4 | 8 KiB | 125 | 125 | 15 ms × (125/125) = 15 ms | 125/0.015s ≈ 8,333 | 8 KiB/s × 8,333 ≈ 65 MiB/s | Compared with Solution 1, the read and write block size and concurrency of Solution 4 are increased, and the throughput and latency performance are improved. The throughput can reach 1 MiB/s. The overall task latency is the lowest among the four solutions, and the IOPS is low. |
Performance of NAS file systems
Item | General-purpose NAS file system | Extreme NAS file system | |||
Storage class | Capacity | Premium | Performance | Standard | Advanced |
Maximum number of files | 1 billion | 1 billion | 1 billion | 0.5 billion | 0.5 billion |
Storage capacity | 0 to 10 PiB | 0 to 1 PiB | 0 to 1 PiB | 100 GiB to 256 TiB | 100 GiB to 256 TiB |
Average latency for reading 4 KiB files from a single-socket server | 10 ms | 2 ms | 2 ms | 1.2 ms | 0.3 ms |
Average latency for writing 4 KiB files to a single-socket server | 10 ms | 2 ms | 2 ms | 1.2 ms | 0.3 ms |
Maximum read throughput | 10 GB/s | 20 GB/s | 20 GB/s | Read and write: 1.2 GB/s | Read and write: 4 GB/s |
Maximum write throughput | 5 GB/s | 5 GB/s | 5 GB/s | ||
Throughput |
|
|
| The maximum throughput is 1,200 MB/s. The throughput increases as the storage capacity of the file system increases, as shown in the following list:
| The maximum throughput is 4,000 MB/s. The throughput increases as the storage capacity of the file system increases, as shown in the following list:
|
IOPS | Upper limit: 15,000 | Upper limit: 30,000 | Upper limit: 30,000 |
|
|
Elastic scale-in | ✓ | ✓ | ✓ | × | × |
Elastic scale-out | ✓ | ✓ | ✓ | ✓ | ✓ |
Scaling step size | 4 KiB | 4 KiB | 4 KiB | 1 GiB | 1 GiB |
Scaling method | Auto scaling | Auto scaling | Auto scaling | Manual scaling | Manual scaling |
References