Docs

Elastic Compute Service

Elastic Block Storage

Elastic remote volumes

Elastic remote volumes are high-throughput, high-IOPS, cost-effective storage solutions for temporary data. This topic introduces the use cases, benefits, and performance of these volumes.

tip

This feature is currently in closed beta. Please contact your account manager for access.

Benefits

Flexibility: Local disks have a fixed capacity. Increasing capacity requires purchasing additional computing resources. These resources can only be created and released with instances. However, elastic remote volumes can be created and detached independently. Volume capacity is customizable and supports elastic scaling.

Cost-effectiveness: Elastic remote volumes offer excellent performance, providing up to 160,000 IOPS and 4 GB/s of throughput per volume. They are more affordable than ESSD FlexPL and TSSD volumes.

Use cases

Although elastic remote volumes are fast and have high IOPS and low latency and cost, they offer no guarantee of data persistence in the event of a failure. Elastic remote volumes are useful in the following cases:

Storing temporary data: Elastic remote volumes can be used for storing temporary data, such as intermediate computation results and cached data. This improves task efficiency, making them ideal for big data processing workloads like Spark.

I/O-intensive scenarios with application-level high availability: Elastic remote volumes deliver exceptional I/O performance, accelerating data read and write efficiency. This feature is ideal for workloads such as relational databases and Elasticsearch.

Volume performance

Performance metrics	Value
Maximum IOPS per volume	160,000
Formula to calculate IOPS per volume	Read: min{50 × Capacity (GiB), 160,000} Write: min{20 × Capacity (GiB), 160,000}
Maximum throughput per volume (MB/s)	4,096 MB/s
Formula to calculate throughput per volume	Read: min{0.8 × Capacity (GiB), 4,096} Write: min{0.4 × Capacity (GiB), 2,048}
Average latency of single-path random writes (block size = 4K)	Sub-millisecond
Maximum capacity per volume (GiB)	System volume: Not supported yet Data volume: 256–16,384 Expansion step size: 1 GiB
Data reliability	The designed failure rate for elastic remote volumes falls between that of local disks and other EBS volume types. Although Volcano Engine uses various technologies to lower failure rates, data persistence is not guaranteed in the event of a failure.
Service availability	If a failure occurs, you can submit a ticket to request manual replacement of the volume.
API name	RSSD_RL0

Billing notes

Elastic remote volumes are billed based on their capacity. Billing begins when you create an elastic remote volume. Both pay-as-you-go and subscription billing options are available. For more information, see Volume billing.

Billing method	Formula	Unit price
Pay-as-you-go	Cost = Unit price × Provisioned storage capacity × Usage time. The billing cycle is accurate to the second.	0.00083 ¥/GiB/hour
Subscription	Cost = Unit price × Provisioned storage capacity × Subscription duration (measured in months).	0.4 ¥/GiB/month

Usage notes

Elastic remote volumes offer no guarantee of data persistence in the event of a failure. To minimize the risk of failure and isolate volumes, we recommend using elastic remote volumes with anti-affinity enabled. Using storage placement groups, EBS places volumes with anti-affinity relationships into separate clusters, thereby isolating the data and protecting it from single-cluster failures.

First, you can create a storage placement group. Then, specify the storage placement group and sub-placement group when creating an elastic remote volume. Volumes with different sub-placement group IDs are deployed in different clusters. For more information, see Managing placement groups.

Notes

Category	Considerations
Instance specifications	You can attach elastic remote volumes to the following instance types: Intel: ecs.c3i, ecs.c3il, ecs.c4i, ecs.c4ie, ecs.c4il, ecs.c3id, and ecs.c4id ecs.g3i, ecs.g3il, ecs.g4i, ecs.g4ie, ecs.g4il, ecs.g3id, and ecs.g4id ecs.r3i, ecs.r3il, ecs.r4i, ecs.r4ie, ecs.r4il, and ecs.r3id AMD: ecs.c3a, ecs.c3al, ecs.c4al, ecs.c4ale, and ecs.c4als ecs.g3a, ecs.g4al, ecs.g4ale, and ecs.g4al ecs.r3a, ecs.r3al, ecs.r4al, and ecs.r4ale
Supported regions	Elastic remote volumes are available in the following regions: China (Beijing) Availability Zone A/C/D China (Shanghai) Availability Zone A/B/C
Volume type	You cannot use an elastic remote volume as a system volume.
Volume performance	Elastic remote volumes do not support additional performance.
Unsupported operations	Elastic remote volumes support the independent creation, detachment, expansion of capacity, and deletion of volumes. However, they currently do not support the following operations: Create snapshots for a volume Create new volumes from snapshots Roll back volumes using snapshots Change the volume type Reinitialize a volume

Creating an elastic remote volume

In the top menu bar, select the project and region in which you want to create a volume.

In the left navigation pane, click "Volumes".

On the "Volumes" page, click the "Elastic remote volume" tab. Then, click "Create elastic remote volume".

On the "Create elastic remote volume" page, configure the volume parameters. For more information about these parameters, see Creating a volume.

After you have finished configuring, click "Confirm order" to submit your volume configurations, the number of volumes you wish to purchase, and the applicable fees.

After confirming that everything is correct, check the box that says, "I have read and agree to the Elastic Compute Service Terms of Service".

Click "Create volume" to submit your order. You will then be redirected to a page where you can check if the volume was created successfully.

Last updated: 2026.05.11 18:51:23