This topic describes the performance metrics and specifications of various Elastic Block Storage (EBS) devices, including cloud disks and local disks.
Performance metrics
The key metrics used to measure the performance of EBS devices include IOPS, throughput, and latency. Some EBS devices also have capacity requirements. For example, enhanced SSDs (ESSDs) at different performance levels (PLs) have different capacity ranges.
- IOPS
IOPS measures the number of read/write operations that can be performed per second. High IOPS is critical for transaction-intensive applications such as database applications. A standard SSD can deliver the committed IOPS only when it is attached to an I/O optimized instance. For information about whether an instance is an I/O optimized instance, see Instance family. The following table describes the common IOPS metrics.
Metric Description Data access method Total IOPS The total number of I/O operations per second. Access locations on storage devices in a continuous or non-continuous manner. Random read IOPS The average number of random read I/O operations per second. Access locations on storage devices in a non-continuous manner. Random write IOPS The average number of random write I/O operations per second. Sequential read IOPS The average number of sequential read I/O operations per second. Access locations on storage devices in a continuous manner. Sequential write IOPS The average number of sequential write I/O operations per second. - Throughput
Throughput measures the amount of data transferred per second. Unit: MB/s. High throughput is critical for applications such as Hadoop offline computing applications that require a large number of sequential read/write operations.
- Latency
Latency measures the amount of time required for an EBS device to process an I/O request. Unit: seconds, milliseconds, or microseconds. High latency may cause performance to degrade or lead to errors in applications that require low latency.
- For latency-sensitive applications such as database applications, we recommend that you use ESSD AutoPL disks, ESSDs, standard SSDs,or local SSDs.
- For applications such as Hadoop offline computing applications that require high throughput but not low latency, we recommend that you use Elastic Compute Service (ECS) instances of the d1 or d1ne instance family that have local SATA HDDs attached.
- Capacity
Capacity is the amount of storage space. Unit: TiB, GiB, MiB, or KiB. EBS capacity is measured in binary units. For example, 1 GiB equals 1,024 MiB.
You cannot use capacity as a metric to measure the performance of EBS devices. However, the performance of EBS devices varies based on their capacity. The larger the capacity of an EBS device is, the stronger its processing capabilities are. EBS devices of the same category have the same I/O performance per unit capacity. However, the performance of a cloud disk linearly increases with its capacity up to the single-disk maximum performance of the disk category. ESSDs of different capacity ranges have different PLs.
For more information about how to test the performance of different categories of EBS devices, see Test the performance of EBS devices and Test the IOPS performance of an ESSD.
Performance of cloud disks
Item | ESSD AutoPL | ESSD | Standard SSD | Ultra disk | Basic disk ③ | |||
---|---|---|---|---|---|---|---|---|
PL | An ESSD AutoPL disk can decouple capacity from performance and deliver a baseline performance equal to that of an ESSD PL1 disk. You can also configure provisioned performance and burst performance for the ESSD AutoPL disk. | PL3 | PL2 | PL1 | PL0 | N/A | N/A | N/A |
Single-disk capacity range (GiB) | 40~32,768 | 1,261~32,768 | 461~32,768 | 20~32,768 | 40~32,768 | 20~32,768 | 20~32,768 | 5~2,000 |
Maximum IOPS | 1,000,000 | 1,000,000 | 100,000 | 50,000 | 10,000 | 25,000 ① | 5,000 | Several hundreds |
Maximum throughput (MB/s) | 4,096 | 4,000 | 750 | 350 | 180 | 300 ① | 140 | 30~40 |
Formula used to calculate the IOPS of a single disk ② |
|
min{1,800 + 50 × Capacity, 1,000,000} | min{1,800 + 50 × Capacity, 100,000} | min{1800 + 50 × Capacity, 50,000} | min{1,800+12 × Capacity, 10,000} | min{1,800 + 30 × Capacity, 25,000} | min{1,800 + 8 × Capacity, 5,000} | N/A |
Formula used to calculate single-disk throughput (MB/s) ② |
|
min{120 + 0.5 × Capacity, 4,000} | min{120 + 0.5 × Capacity, 750} | min{120 + 0.5 × Capacity, 350} | min{100 + 0.25 × Capacity, 180} | min{120 + 0.5 × Capacity, 300} | min{100 + 0.15 × Capacity, 140} | N/A |
Data durability | 99.9999999% | 99.9999999% | 99.9999999% | 99.9999999% | 99.9999999% | 99.9999999% | 99.9999999% | 99.9999999% |
Average single-channel random write latency in milliseconds (block size = 4 K) | 0.2 | 0.2 | 0.2 | 0.2 | 0.3~0.5 | 0.5~2 | 1~3 | 5~10 |
- ① The performance of standard SSDs varies based on the sizes of data blocks. Smaller
data blocks have lower throughput and higher IOPS, as described in the following table.
Data block size (KiB) Maximum IOPS Throughput (MB/s) 4 Approximately 25,000 Approximately 100 16 Approximately 17,200 Approximately 260 32 Approximately 9,600 Approximately 300 64 Approximately 4,800 Approximately 300 - ② In the following examples, a standard SSD is used to describe how to calculate single-disk
performance:
- Single-disk maximum IOPS: The baseline IOPS is 1,800. The IOPS increases by 30 per additional GiB until it reaches the maximum IOPS 25,000.
- Single-disk maximum throughput: The baseline throughput is 120 MB/s. The throughput increases by 0.5 MB/s per additional GiB until it reaches the maximum throughput 300 MB/s.
- ③ Basic disks are cloud disks of a previous generation and are no longer available for purchase.
Performance of local disks
For information about the performance of local Non-Volatile Memory Express (NVMe) SSDs and SATA HDDs, see Local disks.