This sort of testing is done mainly after completion of the software build phase. Minor changes are fixed, entire functionality and minor bugs are being addressed. Eventually goal is about making the core functionality correct or right. In case sanity fails, then entire software build fails. When numerous functionalities are being attached together then small failures are possible.
The objective of this software testing is to make sure than developer has been right in his or her approach. If any small test case fails then there is no way that we can go for the large scale of testing with some higher level of arithmetic. In the beginning phase, if rationality is being checked properly then there is no requirement for the next phase of testing.
You would hardly find sanity testing being documented. Smoke testing is kept with documented while sanity testing is completely different. It can be also considered that this can be a part of the regression testing. Only some critical areas being checked in the sanity portion of software testing. It is also said that this is a process is actually like a health check up campaign for the software applications.
