Reliability of a system is a measure of how resistant it is to failures of various types, for example, attacks, disk or network. Pick ONE OR MORE options.
a) Some companies disable parts of their system on purpose even in production. It is an extreme strategy. They make the system experience failures often and so have proof it can handle them well.
b) Indexing is one of the fundamental approaches to making a system resistant to node failures.
c) Redundancy can help mitigate node failures in the system.
d) Some of the failures can be classified as either network failures, crash failures, or Byzantine failures. The most difficult of these are Byzantine failures.