A regional Enterprise SSD (ESSD) is a new category of ESSD. Data written to a regional ESSD is automatically stored across multiple zones that are physically isolated from each other by data centers, racks, and power supplies. If a physical failure occurs in a zone, a regional ESSD can continue to provide read and write services to ensure business continuity. This topic describes the specifications, billing, limits, and common operations of regional ESSDs.
Introduction
Advantages
Regional ESSDs provide the following advantages over other disk products:
You do not need to manage the complex data replication logic of traditional offline storage. Regional ESSDs rely on physical replication to synchronously write data to multiple zones. This achieves a recovery point objective (RPO) of 0.
For application-level disaster recovery, you can use regional ESSDs to achieve better business continuity without purchasing additional Elastic Computing Service (ECS) instances, bandwidth, or computing resources.
Regional ESSDs provide the same enterprise-level features as other ESSDs.
Data on regional ESSDs is synchronously written to multiple zones. This results in a higher average write latency than PL1 ESSDs. Therefore, regional ESSDs are an ideal choice if your business cannot tolerate data loss from a zone failure and is not sensitive to write latency.
Scenarios
Multi-zone disaster recovery for databases
Traditional database deployment solutions use primary-secondary replication for cross-zone high availability and disaster recovery. These solutions have issues such as replication delay and data inconsistency between the primary and secondary databases. A regional ESSD deployment provides a low-cost alternative. You only need to deploy a compute node in Zone A and attach a regional ESSD, without deploying a compute node in Zone B or configuring primary-secondary replication. The physical replication of the regional ESSD provides data redundancy across zones. If a failure occurs, you can launch a compute node in Zone B to resume services. This can reduce storage costs by 25% and computing costs by 50%.
Cross-zone container deployment
Cross-zone elasticity and disaster recovery for stateful applications remain a major challenge for container deployments. Regional ESSDs allow you to upgrade single-zone stateful applications to provide zone-level disaster recovery capabilities at no transformation cost. If a compute node or a zone fails, or if resources in a single zone are insufficient, you can migrate the container to another zone without complex data synchronization and validation.
Self-built or cloud-deployed Software as a Service (SaaS)
Building or deploying Software as a Service (SaaS) on the cloud typically requires you to set up two ECS clusters in two zones. Regional ESSDs provide a low-cost solution for cross-zone deployments.
Billing
Regional ESSDs are billed based on capacity and support both subscription and pay-as-you-go billing methods. For more information about the billing rules, see Elastic Block Storage billing.
Limits
Region limits
Regional ESSDs are available only in the following regions: China (Hangzhou), China (Shanghai), China (Beijing), China (Zhangjiakou), China (Shenzhen), China (Hong Kong), Singapore, China (Ulanqab), Malaysia (Kuala Lumpur), and Indonesia (Jakarta).
Instance type limits
For information about the instance families that support regional ESSDs, see Instance families.
Feature limits
Feature category | Feature | Is regional Enterprise SSD (ESSD) supported? |
Basic disk features | Create, view, modify, and release | Yes |
Data encryption | Encryption | Yes |
Shared disk attachment | Multi-attach | Yes |
Asynchronous disk replication | Asynchronous replication |
|
Data protection | Create snapshots/Snapshot instant access | Yes |
Automatic snapshot policy | Yes | |
Create a disk from a snapshot | Yes | |
Snapshot-consistent group | No | |
Disk operations | Initialize a disk | Yes |
Resize a disk | Yes | |
Roll back a disk from a snapshot | Yes | |
Use as a system disk | No Note Regional ESSDs can be used only as data disks. | |
Attach a disk | Yes | |
Disk billing | Change the billing method of a disk | Yes |
Disk category | Change the disk category | Yes |
Performance elasticity | Performance provisioning | No |
Performance burst | No |
Disk performance
The following table lists the specifications of regional ESSDs.
Performance Category | Description |
Capacity range (GiB) | 10 to 65,536 |
Maximum IOPS (input/output operations per second) per disk | 50,000 |
Maximum I/O size (KB) | 16 |
Maximum throughput per disk (MB/s) | 350 |
Average latency of random writes per connection (ms) | Millisecond level① |
IOPS formula per disk (baseline performance) | min{1,800 + 50 × Capacity, 50,000} |
Throughput formula per disk (baseline performance, MB/s) | min{120 + 0.5 × Capacity, 350} |
①Latency varies based on regions and zones. You can test the average write latency of a regional ESSD. For more information, see Test the performance of an Elastic Block Storage device.
Use regional ESSDs
Create a regional ESSD
Create a regional ESSD with an instance
Go to the instance purchase page.
Click the Custom Launch tab.
Configure parameters such as the billing method, region, instance type, and image. Pay attention to the following parameters:
Select a region that supports regional ESSDs.
In the Storage section, set Data Disk to Regional ESSD and configure the disk capacity.
For more information about other parameters, see Create an instance on the Custom Launch tab.
Create separately
Go to ECS console - Block Storage.
In the top navigation bar, select the region and resource group of the resource that you want to manage.
Click Create Disk.
Configure the parameters.
Select a region that supports regional ESSDs.
Set Disk Category to Regional ESSD and configure the disk capacity.
For more information about other parameters, see Create a data disk.
Attach the regional ESSD to an ECS instance and initialize it before use.
For more information, see Attach a data disk.
Force-attach a regional ESSD
If a data center or ECS instance fails, you may not be able to detach a regional ESSD. In this case, you can force-attach the regional ESSD to another ECS instance in the same region without detaching the disk first. This speeds up service recovery.
The force-attach operation is available only for regional ESSDs. For other disk categories, you must detach the disk before you can attach it to another instance.
Go to ECS console - Block Storage.
In the top navigation bar, select the region and resource group of the resource that you want to manage.
Find the disk that you want to manage and in the Actions column, click Attach.
Select the target instance and release behavior. Select I confirm that I want to switch the instance to which the disk is attached using the force attach feature. Then, follow the on-screen instructions to attach and initialize the disk.
If data in the memory of the original instance has not been written to the regional ESSD, I/O requests from the original instance to the disk are rejected and fail after you force-attach the disk.
References
For information about how to change the disk category between regional ESSD and other disk types for performance tuning or capacity expansion, see Change the category of a disk.