They were introduced by Azumaya (1951), who named them after Kurt Hensel.
Azumaya originally allowed Henselian rings to be non-commutative, but most authors now restrict them to be commutative.
This should be compared to the fact that for any Zariski open covering
In fact, this property characterises Henselian rings, resp.
For any local ring A there is a universal Henselian ring B generated by A, called the Henselization of A, introduced by Nagata (1953), such that any local homomorphism from A to a Henselian ring can be extended uniquely to B.
The Henselization of A has the same completion and residue field as A and is a flat module over A.
If A is Noetherian, reduced, normal, regular, or excellent then so is its Henselization.
For example, the Henselization of the ring of polynomials k[x,y,...] localized at the point (0,0,...) is the ring of algebraic formal power series (the formal power series satisfying an algebraic equation).
The strict Henselization is not quite universal: it is unique, but only up to non-unique isomorphism.
More precisely it depends on the choice of a separable algebraic closure of the residue field of A, and automorphisms of this separable algebraic closure correspond to automorphisms of the corresponding strict Henselization.
For example, a strict Henselization of the field of p-adic numbers is given by the maximal unramified extension, generated by all roots of unity of order prime to p. It is not "universal" as it has non-trivial automorphisms.