Hi, everyone, and welcome to this lecture. In this lecture, I'm going to explain what snapshots are and what they're used for. On Google Cloud Platform you can use snapshots to backup zonal persistent disks and regional persistent disks. A snapshot is an exact copy of a disk as it exists at a specific point in time and it can even be taken while reads and writes are happening to the disk being snapshotted. 

Snapshotting a disk is like stopping the disk head while the disk is in use and then copying all the sectors from the disk. Because snapshots are so easy to use, they're an appealing backup solution for VM instances. By scheduling snapshots on a regular basis, you can ensure you never have to deal with lost data.

When working with snapshots, it's important to note that they are incremental and they are automatically compressed. What this means is that the first successful snapshot of a persistent disk is, in fact, a full snapshot. It contains all the data that resides on the persistent disk being snapshotted. However, the second snapshot only contains new data that's been written to the disk since the first snapshot. The second snapshot will also contain existing data that's been modified since the first snapshot. Although the second snapshot doesn't contain any data from the first, the second snapshot does contain pointer references to the first snapshot for any unchanged data that's contained in the first snapshot.

If you take a third snapshot, that third snapshot will contain any new or changed data since the second snapshot. Instead of including unchanged data from the first two snapshots, the third one will contain pointer references to blocks in the first two snapshots for any unchanged data. This process repeats itself for any subsequent snapshots that you take.