Escape analysis is a code optimization that determines how wide the scope for a particular object is, and potentially removes that object. If it can be proven that an object only exists in a finite scope and doesn't "escape" from it, for example by being passed as a parameter to method calls within the scope, the object need not be allocated and can be represented as its fields as local variables instead. This saves allocation overhead. This is equivalent to allocating an object on the stack instead of on the heap in languages like C++.