Local disks are disks that are attached to the same physical machine that hosts their ECS instance. Local disks provide local storage and access for ECS instances. Local disks are cost-effective and provide high random IOPS, high throughput, and low latency.

Disk categories

Note This topic provides information about the local disks that are sold together with ECS instances. For more information about the performance of instance families using local SSDs and big data instance families, see Instance families.

Local disks are suitable for business scenarios that require high storage I/O performance, massive storage, and high cost-effective performance. Alibaba Cloud provides the following two categories of local disks:

  • NVMe SSDs

    Instances of the following instance families are attached with NVMe SSDs:

    • Instance families with local SSDs: i2, i2g, i2ne, i2gne, and i1
    • Compute optimized instance family with GPU capabilities: gn5
    • Visualization and compute optimized instance family with GPU capabilities: ga1
    For example, instance families with local NVMe SSDs are suitable for the following scenarios:
    • Services such as online gaming, e-commerce, live streaming, and media services that run I/O-intensive applications and require high I/O performance and low latency.
    • Services that require high I/O performance and a high-availability architecture at the application layer, such as NoSQL databases (including Cassandra, MongoDB, and HBase), MPP data warehouses, and distributed file systems.
  • SATA HDDs

    Instances of the d1ne and d1 instance families are attached with SATA HDDs. These disks are the preferred storage media for industries with high requirements for big data computing, storage, and analysis, such as the Internet and financial industries. The disks are suitable for business scenarios that require massive storage and offline computing. These disks can fulfill the requirements of distributed computing services such as Hadoop, which have high requirements for storage performance, storage capacity, and internal bandwidth.

NVMe SSD performance

  • The following table describes the performance metrics of NVMe SSDs that i2 and i2g instance families use.
    NVMe SSD performance metric Single disk performance Overall instance performance*
    ecs.i2.xlarge and ecs.i2g.2xlarge Other i2 and i2g instance types
    Maximum capacity 894 GiB 1,788 GiB 8 × 1,788 GiB
    Maximum read IOPS 150,000 300,000 1,500,000
    Maximum read throughput 1 GB/s 2 GB/s 16 GB/s
    Maximum write throughput 0.5 GB/s 1 GB/s 8 GB/s
    Access latency Microsecond (μs)

    *Overall instance performance in the table applies only to ecs.i2.16xlarge. This performance represents the highest performance level for local storage in the i2 instance family.

  • The following table describes the performance metrics of NVMe SSDs that i2ne and i2gne instance families use.
    NVMe SSD performance metric ecs.i2ne.xlarge and ecs.i2gne.2xlarge ecs.i2ne.2xlarge and ecs.i2gne.4xlarge ecs.i2ne.4xlarge and ecs.i2gne.8xlarge ecs.i2ne.8xlarge and ecs.i2gne.16xlarge ecs.i2ne.16xlarge
    Maximum capacity 894 GiB 1,788 GiB 2 × 1,788 GiB 4 × 1,788 GiB 8 × 1,788 GiB
    Maximum read IOPS 250,000 500,000 1,000,000 2,000,000 4,000,000
    Maximum read throughput 1.5 GB/s 3 GB/s 6 GB/s 12 GB/s 24 GB/s
    Maximum write throughput 1 GB/s 2 GB/s 4 GB/s 8 GB/s 16 GB/s
    Access latency Microsecond (μs)
    Note We recommend you use the latest version of Alibaba Cloud Linux 2 to ensure maximum throughput performance of disks for Linux instances. Otherwise, the maximum IOPS may not be available for Linux instances.
  • The following table describes the performance metrics of NVMe SSDs that the i1 instance family uses.
    NVMe SSD performance metric Single disk performance Overall instance performance***
    Maximum capacity 1,456 GiB 2,912 GiB
    Maximum IOPS 240,000 480,000
    Write IOPS** min{165 × Capacity, 240000} 2 × min{165 × Capacity, 240000}
    Read IOPS**
    Maximum read throughput 2 GB/s 4 GB/s
    Read throughput** min{1.4 × Capacity, 2000} MB/s 2 × min{1.4 × Capacity, 2000} MB/s
    Maximum write throughput 1.2 GB/s 2.4 GB/s
    Write throughput** min{0.85 × Capacity, 1200} MB/s 2 × min{0.85 × Capacity, 1200} MB/s
    Access latency Microsecond (μs)
    **The performance of a single NVMe SSD is calculated as follows:
    • Write IOPS: 165 IOPS per GiB, up to 240,000 IOPS
    • Write throughput: 0.85 MB/s per GiB, up to 1,200 MB/s

    ***Overall instance performance in the table applies only to ecs.i1.14xlarge. This performance represents the highest performance level for local storage in the i1 instance family.

Note To obtain the standard performance data and measure the Quality of Service (QoS) of Alibaba Cloud local disks, you can test the bandwidth, IOPS, and latency of NVMe SSDs. For more information, see Performance tests on Block Storage.

SATA HDD performance

The following table describes the performance metrics of SATA HDDs.

SATA HDD performance metric Single disk performance Overall instance performance****
Maximum capacity 5,500 GiB 154,000 GiB
Maximum throughput 190 MB/s 5,320 MB/s
Access latency Millisecond (ms)

****Overall instance performance in the table applies only to ecs.d1.14xlarge and ecs.d1ne.14xlarge. This performance represents the highest performance level for local storage in the d1 and d1ne instance families.

Billing methods

Local disks are billed along with the instances to which they are attached. For more information, see Subscription and Pay-as-you-go.

Instructions

  • Local disks are attached to a single physical server, which increases the risk of single point of failure (SPOF). The reliability of data stored on local disks depends on the reliability of the physical server.
    Warning For example, data stored on local disks may be lost when a hardware failure occurs. We recommend that you do not use local disks for long-term data storage.
    • To ensure data availability, we recommend that you implement data redundancy at the application layer. You can use deployment sets to distribute ECS instances across multiple physical machines to achieve high availability and disaster recovery. For more information, see Create a deployment set.
    • If your applications do not have data reliability architecture, we recommend that you use cloud disks or the backup service in your ECS instances for data reliability. For more information, see Cloud disks overview or What is Hybrid Backup Recovery?.
  • After you purchase an ECS instance that is attached with a local disk, you must log on to the instance to partition and format the local disk. For more information, see Partition and format the local disk.
  • The following operations are not supported on a local disk:
    • Create a separate local disk.
    • Use a snapshot to create a local disk.
    • Attach a local disk.
    • Detach and release a local disk separately.
    • Resize a local disk.
    • Reinitialize a local disk.
    • Create a snapshot for a local disk.
    • Use a snapshot to roll back a local disk.

Disk initialization sequence

When you create an ECS instance that is equipped with local disks, all disks are initialized based on the following rules:
  • Rule 1: If the specified image does not have data disk snapshots, the local disks are initialized prior to the cloud disks that are created along with the ECS instance.
  • Rule 2: If the specified image has data disk snapshots, the initialization sequence of the data disks corresponds to the sequence of data disks retained in the snapshots. The initialization sequence of the remaining disks is subject to Rule 1.
For example, for a Linux image that contains snapshots of two data disks, the disks are initialized in the following sequence:
  • If the original device names of the two data disks are /dev/xvdb and /dev/xvdc, Alibaba Cloud first allocates /dev/xvdb and /dev/xvdc to specified data disks in the image. The system disk is initialized first. Then, the initialization proceeds in the following sequence: data disk 1 specified in the image, data disk 2 specified in the image, local disk 1, local disk 2, cloud disk 1, cloud disk 2, and cloud disk N, as shown in the following figure.Rule 2 - Schematic 1
  • If the original device names of the two data disks are /dev/xvdc and /dev/xvdd, Alibaba Cloud first allocates /dev/xvdc and /dev/xvdd to specified data disks in the image. The remaining device names are preferentially allocated to the local disks. The system disk is initialized first. Then, the initialization proceeds in the following sequence: local disk 1, data disk 1 specified in the image, data disk 2 specified in the image, local disk 2, cloud disk 1, cloud disk 2, and cloud disk N, as shown in the following figure.Rule 2 - Schematic 2

Lifecycle

A local disk has the same lifecycle as the instance to which it is attached. For more information, see ECS instance lifecycle.

Impacts of instance operations on the data stored on local disks

The following table describes the impacts of instance operations on the data stored on local disks.

Instance operation Data stored on the local disk Local disk
Restart the operating system, restart an instance by using the ECS console, or forcibly restart an instance. Retained Retained
Shut down the operating system, stop an instance by using the ECS console, or forcibly stop an instance. Retained Retained
Automatically restore an instance. Erased Released
Release an instance. Erased Released
A subscription instance is stopped when it expires or a pay-as-you-go instance is stopped when there is an overdue payment. However, the instance has not been released. Retained Retained
A subscription instance is stopped when it expires or a pay-as-you-go instance is stopped when there is an overdue payment. After that, the instance is released. Erased Released
Renew an expired subscription instance. Retained Retained
Reactivate a pay-as-you-go instance that is stopped due to an overdue payment. Retained Retained

References

For more information about phased-out local SSDs, see Ephemeral SSDs.