This topic describes the architecture and terms of File Storage NAS (NAS). To better understand and use NAS, we recommend that you read this topic.
Architecture of NAS
NAS is a distributed file sharing solution that provides high scalability, reliability, and performance. NAS allows you to migrate business data from on-premises storage systems to Alibaba Cloud without the need to modify your application code. Each NAS file system can be mounted on thousands of compute nodes, such as Elastic Compute Service (ECS) instances and Container Service for Kubernetes (ACK) clusters. The following figure shows the relationships between compute nodes and a NAS file system.
Alibaba Cloud terms
Term | Definition | Usage notes |
UID | The ID of the Alibaba Cloud account. The account ID is a string of numbers. You can view the account ID on the Account Management page of the Alibaba Cloud Management Console. |
|
A geographic location where Alibaba Cloud data centers are deployed. Examples: China (Hangzhou) and China (Beijing). | ||
An isolated location in a region where the power supply and network are independent. Examples: Hangzhou Zone B and Frankfurt Zone A. |
| |
A logically isolated network. You can connect two VPCs by using CEN. | A network environment is a prerequisite for compute nodes to access NAS file systems. We recommend that you use a VPC rather than the classic network for better user experience. |
NAS-specific terms
Term | Definition | Usage notes |
A place where files are stored. A file system can be mounted on a compute node by using a mount target. The data and metadata in the file system are stored in the traditional directory tree structure. A file system supports simultaneous access from hundreds of compute nodes for data sharing and highly concurrent read and write operations. | NAS file systems are ideal file storage applications rather than network disks. NAS file systems cannot be accessed by using RESTful APIs. | |
A billing method that is used to offset the storage usage fees of General-purpose NAS file systems. This billing method allows you to plan your NAS resources and enjoy discounts. Important You can no longer purchase storage plans. However, you can renew or upgrade existing storage plans. |
| |
The process to connect a compute node to a NAS file system. After a NAS file system is mounted on a compute node, the compute node can access data in the file system. |
| |
An endpoint of a NAS file system in the classic network or in a VPC. Each mount target is displayed as a domain name. |
| |
An access method provided by NAS for different applications and users to access NAS file systems. Access points allow you to easily manage the access of different applications and users to datasets. |
| |
A group that defines the permissions to access a file system by using mount targets. You can add rules to a permission group to specify the authorized IP addresses and the corresponding read and write permissions on the file system. |
| |
A storage class of General-purpose NAS file systems. Capacity NAS file systems are suitable for cost-sensitive file sharing workloads that require moderate latency, such as database backup, log storage, Windows user directory, and Linux home directory. | The storage capacity of a Capacity NAS file system can be automatically scaled. The maximum storage capacity is 10 PiB. | |
A storage class of General-purpose NAS file systems. Premium NAS file systems are suitable for latency-sensitive file sharing workloads that require low latency, such as container data persistence, AI training data storage, manufacturing simulation, and genetic computing. | The storage capacity of a Premium NAS file system can be automatically scaled. The maximum storage capacity is 1 PiB. | |
A storage class of General-purpose NAS file systems. Performance NAS file systems are suitable for latency-sensitive file sharing workloads that require low latency, such as Linux or Windows applications for enterprises, container persistent volumes (PVs), web content management, and genetic computing. | The storage capacity of a Performance NAS file system can be automatically scaled. The maximum storage capacity is 1 PiB. | |
A regular data storage pool for General-purpose NAS file systems. | After you create a General-purpose NAS file system (Capacity, Premium, or Performance), data is stored in the Standard storage class. If you want to dump data to the Infrequent Access (IA) or Archive storage class, you must configure a lifecycle policy for the file system. | |
A low-cost data storage pool for General-purpose NAS file systems. The IA storage class is suitable for storing data that is accessed less than three times each month. |
| |
An ultra-low-cost data storage pool for General-purpose NAS file systems. The Archive storage class is suitable for storing data that is accessed less than twice each quarter. |
| |
A software protocol that is used to mount a NAS file system on a compute node. The NFS protocol is applicable to Linux clients and most containers. |
| |
SMB | A software protocol that is used to mount a NAS file system on a compute node. The SMB protocol is applicable to Windows clients and Windows containers. | None. |
POSIX-based APIs | A set of standard interfaces that are used to access files in a Linux operating system. You can call the POSIX-based APIs by using a programming language in applications to manage files. | Linux applications also use POSIX interfaces to manage on-premises files. Therefore, if you migrate on-premises files to a NAS file system, you can access the files without the need to modify the application code. To achieve optimal performance of NAS file systems, we recommend that you increase the concurrency of your applications and increase the block size for read and write operations. |