Updated on 2024-09-30 GMT+08:00

Overview

Introduction

CCE allows you to mount a volume created from a Scalable File Service (SFS) file system to a container to store data persistently. SFS volumes are commonly used in ReadWriteMany scenarios for large-capacity expansion and cost-sensitive services, such as media processing, content management, big data analysis, and workload process analysis. For services with massive volume of small files, SFS Turbo file systems are recommended.

Expandable to petabytes, SFS provides fully hosted shared file storage, highly available and stable to handle data- and bandwidth-intensive applications

  • Standard file protocols: You can mount file systems as volumes to servers, the same as using local directories.
  • Data sharing: The same file system can be mounted to multiple servers, so that data can be shared.
  • Private network: Users can access data only in private networks of data centers.
  • Capacity and performance: The capacity of a single file system is high (PB level) and the performance is excellent (ms-level I/O latency).
  • Use cases: Deployments/StatefulSets in the ReadWriteMany mode and jobs created for high-performance computing (HPC), media processing, content management, web services, big data analysis, and workload process analysis

Performance

CCE supports SFS Capacity-Oriented and SFS 3.0. For details about file system types, see File System Type.
  • SFS Capacity-Oriented file systems are sold out and cannot be used to create PVs on the CCE console. However, you can still create PVs using existing SFS Capacity-Oriented file systems through kubectl.
  • SFS 3.0 is being rolled out in different regions, but it may not be available in all regions. If you encounter any issues, contact SFS customer support or wait for further updates. If the region where your application is located already has SFS 3.0 available, use it for new applications and migrate existing SFS Capacity-Oriented file systems to SFS 3.0 as soon as possible to prevent any service disruptions caused by insufficient capacity.
Table 1 Performance

Parameter

SFS Capacity-Oriented

SFS 3.0

Maximum bandwidth

2 GB/s

1.25 TB/s

Maximum IOPS

2000

Million

Latency

3–20 ms

10 ms

Maximum capacity

4 PB

EB

Application Scenarios

SFS supports the following mounting modes based on application scenarios:

  • Using an Existing SFS File System Through a Static PV: static creation mode, where you use an existing SFS volume to create a PV and then mount storage to the workload through a PVC. This mode applies if the underlying storage is available or billed on a yearly/monthly basis.
  • Using an SFS File System Through a Dynamic PV: dynamic creation mode, in which you do not need to create SFS file systems beforehand. Instead, specify a StorageClass when creating a PVC. Then, a file system and PV will be created automatically. This mode applies to scenarios where no underlying storage is available.

Billing

  • The default billing mode of an SFS volume automatically created using StorageClass is Pay-per-use, indicating that it is billed based on the used storage capacity and duration. For details about SFS pricing, see Billing.
  • To use a yearly/monthly-billed SFS volume, use an existing one.