Snapshots are used to back up data from persistent disks to Cloud Storage, but they are not visible in your buckets as they are managed by the snapshot service. Snapshot is not available for local SSD. They can be used to create new disks, which may be in another region or zone in the same project, and thus they form the basis of VM migration. But you cannot share snapshots across projects.

Snapshots do not back up VM metadata, tags, and so on. In this way, they are different from images—which are used primarily to create instances or configure instance templates. Snapshots can be created from persistent disks even while they are attached to running instances. Images, in turn, can be either public or custom, and are an important tool in creating managed instance groups to clone a given single instance.

Snapshots are incremental and can hence be used to create regular snapshots on a persistent disk faster and at a much lower cost than regularly creating a full image of the disk.

On Windows, you can enable Volume Shadow Service (VSS), which enables a disk to be backed up without having to be shut down. Snapshots are the size of the existing disk. They cannot be restored to a smaller disk or other disk types. Thus, shrinking a disk would be an OS copy process of data from one attached disk to a second smaller attached disk. Multiple copies of each snapshot are redundantly stored across multiple locations with automatic checksums to ensure the integrity of data.