Most cloud providers and storage vendors provide data volumes with specified IO Quality of Service (performance) parameters, such as IOPS or throughput, and adjust them while the workload is running. However current Kubernetes does not support this feature, and this will causes the following problems:

1. As a cluster administrator, you need to manually operate the underlying storage system to modify volume attributes. But this will be very difficult & complicated for the operator.

2. As a storage vendor, while integrating storage with Kubernetes, they may choose to support this feature. But there are many limitations, and it might be inconsistent with the community.

In this proposal we will describe the current practice of some storage providers (such as HwameiStor, AWS, etc.) and propose extensions to the Kubernetes Persistent Volume API to allow users to dynamically control these volume options.