Products



Virtual Machines

Documentation > Virtual Machines > Storage Overview

Product

Storage overview

JD Cloud Instance provides different types and combinable data storage devices. Different types and different capacity sizes of storage devices have different performance and price, which are applicable to different usage scenarios. Depending on the dimensions, they can be divided into the following categories:

Mechanical hard drives System Disk Local disk

Dimensionality	Category	Description
Storage Media	Data is stored on a mechanical hard drive. Good price/performance ratio and better read/write speed.
Storage Media	SSD drives	Data is stored on a solid state drive (SSD). Its IOPS and read/write speed performance are higher than mechanical hard drives, and the price is higher than mechanical hard drives.
Usage Scenarios	A system application for storing operating system data and controlling and managing instances, providing the underlying data environment for instance operation, with the same lifecycle as the instance.
Usage Scenarios	Data Disk	For storing user data, users can purchase on-demand and control the mounting and unmounting by themselves.
Architecture Patterns	A local disk is a storage area divided from the local storage area of the physical machine where the instance is located. If the local disk and the instance are arranged on the same physical machine, the access delay is short, but there is single point failure risk of data. At the same time, the local data disk will be emptied when the instance is configured with no billing during downtime and is disabled. If your application architecture cannot achieve data reliability, it is recommended that you use cloud disk to store important data.
Architecture Patterns	Cloud Disk	Cloud disk is an elastic, highly available and highly reliable data block-level memory. The data in the cloud disk is stored in triple real-time copy mode to avoid data unavailability due to component failure. Elastic expansion is allowed for cloud disk capacity. You can expand data storage space at a low price in a few minutes to achieve persistent storage of your data.

Classification according to architectural patterns

Local Disk

The local disk and the instance are in the same physical machine. Instances using local disks can obtain smaller data access latency, but there is a risk of single-point failure. In addition, the life cycle of the local disk is the same as that of the instance. The data on the local system disk is lost when the instance is deleted. At the same time, the local data disk will be emptied when the instance is configured with no billing during downtime and is disabled. If your application architecture cannot achieve data reliability, it is recommended that you use cloud disk to store important data.

Cloud Disk

Cloud Disk Services are data block-level storage devices with a triple copy architecture that makes them highly available.

JD Cloud provides five types of cloud disks with different storage media, namely, general-purpose SSD cloud disk, performance-type SSD cloud disk, and capacity-type HDD cloud disk. Currently, you may use cloud disk whose life cycle is independent from that of the instance as the data disk of the instance. Such cloud disk can be either purchased along with the instance, or be separately purchased after the creation of the instance and then attached to the instance, and can be unattached from the current instance and attached to other instances.

In addition, it can support the creation of Cloud Disk Services that can be mounted in multiple points. For details, please refer to Multipoint Mount Cloud Disk Service.

You can create a snapshot of the Cloud Disk Service to keep a backup copy of the data and create a new Cloud Disk Service from the snapshot at any time and connect it to another instance. For more information about Cloud Disk Service, please refer to Cloud Disk Service Product Documentation.

Classification according to usage scenarios

System Disk

The instance supports the use of local disk and Cloud Disk Services as its system disk.

If you use the local disk as the system disk to get lower data access latency, 40GB capacity is given away, and the system disk is not supported to expand the capacity operation.

If you use Cloud Disk Service as system disk can get higher data reliability, support configuration 40GB ~ 500GB capacity, support offload after expansion operation.

Notification: The local system disk is now being phased out, so we recommend you to use the Cloud Disk Service as a priority.

Data Disk

The instances support the use of local disk and the Cloud Disk Service as the data disk. Where,

Local Disk Currently only some instance types (e.g.: GPU Standard, Storage Optimized, etc.) are provided, the number of local data disks and the capacity of a single disk depend on the configuration in the specifications, and they have the same life cycle as the instance and do not allow operations to unmount or make capacity resizings. Because of the risk of losing data (e.g., in the event of migration or host downtime), it is not suitable for use in scenarios where there is no data redundancy architecture at the application layer, and is usually used only as a temporary storage medium for intermediate data, and is not recommended for storing important data.
Cloud Disks supports multiple types, including General SSD Cloud Disk, Performance SSD Cloud Disk and Capacity HDD Cloud Disk. Each instance can attach a maximum of 7/8 cloud disks as data disks (if the system disk is a cloud disk, 7 cloud disks can be attached as data disks). In the case of adjustment to this quota, please go to Open Ticket for application. Please refer to JD Cloud Disk Service Documentation for more information.

Block Storage Device Mapping

Each instance has a system disk to ensure the storage of basic operational data, and more cloud disks can be mounted to the instance as data disks, please refer to Mount Cloud Disk for details. Instances use block storage device mapping to map these storage devices to their own recognizable locations, automatically creating block storage device mappings for the local disk and Cloud Disk Services mounted to them.

The following diagram shows how to map a block storage device to an instance, map /dev/vda to the system disk, and map the two data disks to /dev/vdb and /dev/vdc respectively. The device names may vary within different operating systems, detailed rules can be found in Device Name Assignment Rules.

Storage billing

The local system disk is free of charge, with a fixed capacity of 40GB. The local data disk has the same life cycle as the instance, and its price is included in the instance price.
The price of Cloud Disk Service system disk and data disk depends on the disk type and capacity, see Cloud disk price for detailed price.

Multipoint Mount Cloud Disk Service

Multi-point mount Cloud Disk Service is a data block-level storage device that supports multiple instances of concurrent read and write access, with multiple concurrency, high performance, high reliability and other features, and supports high availability architecture scenarios such as Oracle RAC, which is commonly used in government and enterprise, finance and other industries. A single block of shared block storage can be mounted to up to 16 instances simultaneously. For more details, see How to use Multipoint Mounted Cloud Disk Service

Multipoint Mounted Cloud Disk Service can only be used as a data disk and is not supported to be created along with the instance for now. You need to create it separately and specify the multipoint mount attribute when creating it, please refer to Create Cloud Disk Service. Setting the Delete with Instance property is not supported, and the multipoint mounted Cloud Disk Service will be retained when its mounted instance is deleted.

When mounted to an instance, it shares the single instance mountable Cloud Disk Service limit with non-multipoint mountable Cloud Disk Services, i.e., one instance can mount up to 8 Cloud Disk Services.