Let’s say you backup your data on someone else’s computer. If that computer goes down, or its hard drive breaks, you’ll need to get another computer and restore your data in order to have a backup.
If you use a cloud storage platform, though, you don’t need to complete this backup process yourself. The platform has data backup, replication, versioning, and other controls already built into it. With versioning, for example, you can keep a historical record of all the real-time backups of your data. Even people with access to the data can’t delete or overwrite the old versions. As a result, you don’t lose any instance of your data between the time that it was written and the time of the backup.
Cloud providers like Google Cloud, Azure, and AWS can do this because they use data centers and replicate data across several different physical drives and locations. They’ve also developed processes for faulty hardware, disasters, and other potential issues.
All of this improves their data backup, recovery, and reliability capabilities. Even if something happens to data in one of the data centers, the real-time backups across multiple locations minimize downtime and data loss.
A single file stored in S3, for instance, will have at least three different copies in three different physical locations. So you would need at least three of “someone else’s computer” to match this capability—but far more when you also consider the redundancy and continuous integrity checking that is baked into the service.
And that’s only for storage. The number grows when you consider accessing that file. For example, if your system needs to support reading that file 5,000 times per second (which S3 does support without any advance warning), that will require even more computers. That’s what makes the cloud more than just someone else’s computer. Operating on a much larger scale allows it to provide much better capabilities.