@ Identity disadvantages scope_identity () replace

You may be familiar with @ identity in SQL Server. It is known that it is used to obtain the identity value of the last inserted data in the data table.
For example, if Table A has a field with ID auto-increment of 1 and the id value is 100, if I insert a piece of data to table A and
Select @ identity, then it returns the value of 101, the last identity domain (that is, the ID domain.

Now the question is, why should we use @ identity with caution? The reason is that @ identity always obtains the value of the auto-increment field of the last change data,
The scope constraints of the change operation are ignored. For example, I have two tables, table A and Table B. Now I have defined an insert trigger on Table,
When a piece of data is inserted in Table A, a piece of data is automatically inserted in table B. Note that there are two atomic operations: Insert a piece of data in a, and then insert a piece of data in B.

Now, assume that both table A and table B have an identity auto-incrementing domain. After inserting a data entry in Table A, we use
In the select @ identity output, is the output A or B's auto-incrementing Domain value? The answer is obvious, who will output the last insert,
That's B. So I wanted to get the auto-incrementing Domain value of A, and the result obtained the auto-incrementing Domain value of B. A Bug was born along with it.
Will affect the entire system data chaos.

Therefore, we recommend that you use @ identity with caution, and use the scope_identity () function whenever possible. Scope_identity ()
The value of the last auto-increment domain is obtained, but it is only within the range of an operation. Unlike @ identity, it is the last operation to take the global operation.
The value of the auto-increment field generated.