It is predictable that technology will fail or become obsolete over time.
What we ask ourselves is when do we expect a specific piece of technology to fail and what can we do to mitigate the risk of it failing. There are industry studies that give us statistics and timelines for when we should expect failure to happen, the meantime before failure (MTBF).
We configure critical server resources with redundancy to account for individual components failing. And we recommend replacing critical equipment before it hits the upper end of that expected failure window.
We also make and store a second copy of data, called a backup, to mitigate the risk of predictable hardware failure, unpredictable user error, or unexpected data corruption.
Backups are taken using a specialty software, called backup software. This software not only copies the data, but compresses it the data for storage and manages how much backup history is retained.
Local, or on premise, backups are stored on a storage device at the same location as the devices being backed up. When someone speaks of backups, this is the type of backup they mean.
Disaster recovery backups are a set of backups stored offsite at a different, remote location, geographically separated from the devices being backed up. The offsite copy of data mitigates the risk of something happening to your office location. When someone says "disaster recovery", DR, or "cloud" backup this is the type of backup they mean.
In the event of an issue with a company's data storage, being able to access and restore data QUICKLY is critical. DR, offsite, or "cloud" storage is slower to access and restore because it is physically located outside your organization and retrieved over the Internet. This speed to restore difference is why local backups are recommended in combination with DR backups.