NASA, and I'd imagine most space agencies around the world, demands abort systems of its commercial partners capable of destroying a rocket should it venture too close to populated areas. Note that abort system can also refer to launch escape systems, but to keep this question narrow in scope I'm only referring to the ability to blow up a misbehaving rocket.
The engineering here is fascinating, as rockets are basically bombs that have been carefully coaxed into not exploding, but there are circumstances in which you want to ensure reliable and immediate detonation despite having gone to great difficulty to design the rocket to... not do that.
On the technical side, you've got your fuel and oxidizer in two separate tanks, generally cryogenically cooled, often but not always at very high pressure. Given the danger of unburned rocket propellant being sprayed down on an inhabited area, one of the goals of this detonation is to rapidly consume both fuel and oxidizer. The other goal is to separate the rocket into pieces capable of causing less damage than being hit by the rocket as a whole. It seems that the three steps of combining all the fuel and oxidizer, burning them both all at once, and splitting up the rocket, all in one precisely timed explosion of a thing that has been built not to explode, is an engineering marvel in itself, one of the many involved in building rockets.
How do they do it?
It seems that the three steps of combining all the fuel and oxidizer, burning them both all at once, and splitting up the rocket, all in one precisely timed explosion of a thing that has been built not to explode, is an engineering marvel in itselfWhy? Breaking stuff is easy. The hard engineering part is the "making it not explode" bit. To make it explode, all you have to do is compromise (ie : break) one of the hundreds of engineering controls you put in place to prevent it from exploding and the rest happens all by itself. $\endgroup$