replication id

[Mohammadsharif]

What is replication ID ,,please give a clear example ,thanks in advance

[CJ_London]

A Replication ID (which is actually a GUID, or globally unique ID) is guaranteed to be totally unique because it is generated from a number of different bits of information, such as the devices parameters (e.g. the date and time, and the free byte count on the C: drive.
They look like this {8EC7B8E3-3B51-4C0C-8481-7AEEF9074EBE} (note this is not text, but a bit like a date field which is a number formatted to look like a date) and require 16 bytes- a long integer (the autonumber field) is only use 4 bytes.

They are normally used where you know you are going to be merging data on a regular basis (e.g. taking a copy of the db offsite, making additions/changes, then returning and merging the data).

Because of their size, processing can be slower although with modern computers this may not be so apparent. You need to weigh up the benefits of a totally unique ID for ease of merging against potential performance issues. There is no benefit that I can see from using them in a normal db scenario.