Elastic Block Storage (EBS) is a high-performance, low-latency block storage service provided by Alibaba Cloud for Elastic Compute Service (ECS). EBS supports random read and write to meet the data storage requirements of your business. EBS devices can be used as physical hard disks. You can partition and format EBS devices and create file systems for them.
Alibaba Cloud provides a variety of EBS devices for ECS, such as cloud disks based on a distributed storage architecture and local disks located on the same physical machine that hosts the associated ECS instances.
|Cloud disk||ESSD AutoPL disk||You can specify provisioned performance and enable performance burst for ESSD AutoPL
disks based on your business requirements. ESSD AutoPL disks inherit the features
and performance of enhanced SSDs (ESSDs) and provide the capability of decoupling
disk capacity from disk performance.
We recommend that you use ESSD AutoPL disks in the following scenarios:
|ESSD AutoPL disks|
|ESSD||ESSDs utilize a next-generation distributed block storage architecture and provide
25 Gibabit Ethernet and remote direct memory access (RDMA). Each ESSD has low one-way
latency and can deliver up to 1,000,000 random read/write IOPS.
We recommend that you use ESSDs for scenarios such as OLTP databases, NoSQL databases, and ELK distributed logs.
|Standard SSD||Standard SSDs are high-performance disks that offer consistent high random IOPS and
high data reliability.
We recommend that you use standard SSDs for scenarios such as I/O-intensive applications, small and medium-sized relational databases, and NoSQL databases.
|Ultra disk||Ultra disks are cost-effective and offer medium random IOPS and high data reliability.
We recommend that you use ultra disks as system disks or for scenarios such as development and testing.
|Basic disk||Basic disks are the previous generation of disks and are unavailable for purchase.||None|
|Local disk||Local disks are physical disks attached to physical machines that host ECS instances.
Local disks provide local storage for ECS instances. They are applicable to business
scenarios that require high storage I/O performance, a large amount of storage, and
high cost-effectiveness. Local disks provide low latency, high random IOPS and throughput,
and high cost performance.
Note All of the local disks for an instance reside on a single physical machine. This increases the risks of single points of failure (SPOFs). The durability of data stored in a local disk is determined by the reliability of the associated physical machine. 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 for high availability and disaster recovery.
- Data durability during read and write operations
Three copies of your business data are stored in a block storage cluster within the same zone to ensure 99.9999999% (nine 9's) data durability during read and write operations. For more information, see Triplicate storage.
- Proactive backup
You can create snapshots for disks on a regular basis to enhance data security. Snapshots are a backup service provided by Alibaba Cloud.They back up data stored on disks to ensure that data such as logs and customer transactions can be queried. For more information, see Snapshot overview.
- Data erasure mechanism
When you delete a piece of data, the data is completely erased from the distributed block storage system and can no longer be accessed by other users. The following mechanisms ensure that deleted data is completely erased:
The storage system performs sequential writes to append data to existing files at the underlying layer of cloud disks. This mechanism makes full use of high-bandwidth and low-latency sequential writes to physical disks. If you delete a logical space from a cloud disk after data is appended at the underlying layer of the cloud disk, this operation is recorded as metadata. The storage system returns only zero for all requests of reading data from the logical space. Similarly, when you overwrite the data that resides in the logical space of a cloud disk, the storage system does not immediately overwrite the data in the corresponding physical space. Instead, it modifies the mapping between the logical space and the physical space. This ensures that data that has been overwritten is no longer readable. Data fragments that result from delete or overwrite operations are forcefully and permanently deleted from the underlying physical disks.
When an EBS device (cloud disk) is released, the storage system immediately destroys the metadata of the device to ensure that the data is no longer accessible. At the same time, the physical storage space that corresponds to the disk is recycled. The physical storage space must be cleared before it is re-assigned to store data. Before data is written to a new cloud disk, the system returns only zero for all read requests.
- Data encryption
If your applications are data-sensitive, we recommend that you encrypt the storage devices that you use. Cloud disks and their snapshots are encrypted by using keys based on the industry-standard AES-256 algorithm. Data is automatically encrypted when it is transmitted from ECS instances to cloud disks. Encrypted data is automatically decrypted when it is read. For more information, see Encryption overview.
For information about the limits and quotas of EBS devices, see the "Elastic Block Storage (EBS) limits" in Limits.
|Creation phase||Create an empty cloud disk or a cloud disk that contains backup data for an ECS instance.||
|O&M phase||View the basic information and monitoring data of cloud disks.|
|Use the snapshot feature to back up and restore data stored in cloud disks.|
|Upgrade the specifications of cloud disks that do not meet your performance requirements.|
|Resize the storage capacity of cloud disks that do not meet your performance requirements.|
|If you want to restore cloud disks to their initial states, you can initialize the disks.|
|If you want to use a new operating system, you can change the operating system of an instance by replacing the system disk.|
|Release phase||When a cloud disk is no longer needed, you can release it to minimize costs.|