If only one number is configured in the buffer (see Administration of the Number Range Buffer), the procedure for assigning numbers is different to the procedure described in How the Number Range Buffer Works.

With one number in the buffer, the number range server is used each time a number is assigned. The numbers are not put in the buffer. This procedure takes a while longer than when numbers are assigned directly in the database (without buffering). Though, here the advantage of using the number range server, is that the number assignment is written directly to the database (commit). Batch jobs in particular, where a lot of time is lost between two commits, benefit from this procedure. A second process that wants to assign a number from the same interval does not have to wait until the first process has executed a commit operation.

Furthermore, the disadvantage that many numbers are lost if the server stops operating, does not apply here. However, the disadvantage that a break in the numbering occurs from rollbacks, does apply here too.

While the number range server is active and assigning a number, a dummy entry exists in the number range buffer, which you can display in SM56. This entry does not have any meaningful values for From number, To number, Release, Critical number and Buffer to No..