Elastic Volume Service (EVS) is a block storage service that uses distributed storage technology. It provides highly reliable, high-performance storage resources for servers. You can choose different types of EVS disks based on your service requirements to achieve fast data read/write and storage. Similar to physical disks on a local PC, EVS disks must be attached to servers before they can be used. They cannot function independently. You can initialize the attached disk, create a file system on it, and persist data on the disk. In this document, EVS disks are sometimes simply referred to as "disks".
In traditional data centers, expanding and managing storage resources often takes considerable time and incurs high costs. As services grow rapidly, data volumes increase sharply and the demand for storage performance rises continuously. Traditional storage solutions struggle to meet these needs. EVS solves these challenges with elastic scaling, high availability, and flexible management, helping users handle data storage demands with ease. For example, during peak hours, you can quickly expand storage capacity to maintain service continuity and stability.
Benefit | Description |
|---|---|
Rich specifications | EVS offers disks with various specifications to meet performance needs in different scenarios, delivering low latency and high IOPS. |
Elastic scalability | You can expand EVS disk capacity as needed to support service growth. EVS allows capacity expansion without interrupting services. |
High security and reliability |
|
Flexible management | You can manage the lifecycle of EVS disks easily through the management console or APIs, including creating, attaching, detaching, and deleting disks. |
EVS disks are block-level storage devices built on a distributed architecture. They provide persistent, highly reliable data storage for servers. Unlike traditional local disks, EVS disks store data across backend storage clusters and are attached to servers as block devices via network protocols. This design combines the performance of local disks with the scalability of cloud storage.
1. Virtualization layer: from physical disks to logical disks
The core of EVS is storage virtualization. A backend cluster pools large numbers of physical disks (SSDs/HDDs) and abstracts them into a unified logical resource pool using a distributed storage engine. Each EVS disk a user creates is essentially a set of logical data blocks in this pool, decoupled from specific physical hardware.
2. Distributed redundancy to prevent data loss
To ensure reliability, EVS uses redundancy mechanisms such as multi-replica or erasure coding (EC). In the three-replica model, each piece of data is simultaneously written to three physical disks across separate nodes and cabinets. If one node fails, services automatically switch to another node, ensuring durability of up to 99.9999999% (nine nines).
3. Network attachment and block mapping
EVS maps remote logical volumes to the server OS using storage protocols such as SCSI or VBD. Once attached, the OS recognizes the EVS disk as a standard block device. You can then partition, format, and use it like a local disk. Read/write requests are transmitted over the network to the backend storage cluster for processing.
4. Elasticity and snapshots
Because EVS disks are decoupled from physical hardware, they support capacity expansion in seconds, online specifications changes, and rollback via snapshots. Snapshots use incremental pointer technology to record only changed data blocks, saving space and enabling second-level consistent backups.
There are three types of storage available for you to choose from: EVS, Scalable File Service (SFS), and Object Storage Service (OBS). Their differences are described in the following table.
Dimension | SFS | OBS | EVS |
|---|---|---|---|
Concept | SFS provides on-demand high-performance file storage that can be shared by multiple cloud servers. You can use SFS as a remote directory for Windows or Linux servers. | OBS provides massive, secure, reliable, and cost-effective data storage for you to store data of any type and size. | EVS provides cloud servers with scalable block storage that features high reliability and high performance and comes with a variety of specifications. EVS disks are like physical disks on PCs. You can use them in different scenarios to meet various service requirements. |
Data storage logic | Stores files. Data is sorted and displayed in files and folders. | Stores objects. Files are saved directly to OBS. The files automatically generate corresponding system metadata. You can also customize the metadata if needed. | Stores binary data and cannot directly store files. To store files, you need to format the disk with a file system first. |
Access method | SFS file systems need to be mounted to cloud servers through the Network File System (NFS) or Common Internet File System (CIFS) protocol before they can be accessed. A network address must be specified or mapped to a local directory for access. | OBS buckets can be accessed through the Internet or Direct Connect. The bucket address must be specified for access, and transfer protocols HTTP and HTTPS are used. | EVS disks can only be used and accessed from applications after being attached to cloud servers and initialized. |
Application scenarios | High-performance computing, media processing, file sharing, content management, and web services NOTE: Mainly suitable for high-performance computing workloads like gene sequencing and image rendering that require high bandwidth for file sharing. | Big data analytics, static website hosting, online video on demand (VOD), gene sequencing, and intelligent video surveillance | High-performance computing, enterprise critical clustered applications, enterprise application systems, and development and testing NOTE: Mainly suitable for high-performance workloads like industrial design and energy exploration that require high speed and high IOPS for high-performance storage. |
Capacity | PiB-level | EiB-level | TiB-level |
Latency | 3–10 ms | Milliseconds | 1 ms |
IOPS/TPS | 10,000 per file system | Tens of millions | 50,000 per disk |
Bandwidth | GiB/s | TiB/s | MiB/s |
Data sharing | Supported | Supported | Supported |
Remote access | Supported | Supported | Not supported |
Used independently | Supported | Supported | Not supported |
EVS disk performance metrics include:
EVS disk types are classified based on performance. You can select the disk type that best matches your workload requirements. For details, see Disk Types and Performance.
EVS provides a web-based console and HTTPS-based APIs that you can use to access the EVS service.
Use APIs if you need to integrate EVS into a third-party system for secondary development. For details, see Elastic Volume Service API Reference.
Use the console if you do not need to integrate EVS with a third-party system. Sign in to the console and choose Elastic Volume Service from the service list.