Highly Reliable Systems is a generic term describing a computer system that is designed to be more reliable than an 'average' system.  Fault tolerant systems are designed such that the failure of one or more components doesn't result in an overall failure of the system and is often not even noticed by the users.  This is often achieved by building systems with duplicate components that can take over when one fails. For example, fault-tolerant computer storage can be implemented using two disk drives -- the same data is stored on both disk drives.  In the event one of the disk fails, the computer storage system continues to operate without the loss of any data.  Cloud services typically are built using highly reliable systems -- in the event one individual computer or component fails, others take over.  Once in a while, when a Netfix movie pauses, it is because the computer you were connected to failed and another took over.

For more information:

• An introduction by PC Magazine
• A more detailed overview by Microsoft