What happens to outgoing IO operations on a volume during its snapshotting

Stateful Workloads
1

in in Portworx documentation here Kubernetes Snapshots and Backups said:

“Portworx will quiesce IO on all volumes in the group and then trigger the snapshot.”

Could somebody point me to any additional documentation describing what is actually happening with outgoing IO operations to the volume during its snapshotting?
In particular, we are interested to get answers to following questions:

  1. Does creation of VolumeSnapshot causes suspend for outgoing IO write operation to the volume?
  2. What would be max/avg duration of such suspend (in sec.) ?
  3. Does the duration affected by the volume’s data size? What would be the dependency pattern?
  4. How the duration affected by disks type used ?
  5. How the duration affected by network performance in the cluster?