We can see this options if your generic data source is made delta possible. We can define delta for generic data source using 3 different fields.
1) Time stamp
2) Calendar Day
3) Numeric Pointer
The use of safety intervals is to make the system to extract delta records which might missed in last extraction (they might not saved at the time of extraction)
Now we will discuss this with delta field as “Numeric Pointer”
Safety Interval Upper Limit: Let’s assume in the last extraction delta pointer set to 1000(numeric pointer). Now in source system we have 100 new/changed records, so new numeric pointer in source system is 1100. If the safety interval upper limit is set to 10, the selection interval set by the system is 1000 to 1090(system will subtract from new numeric pointer value).
Note: If you use upper limit, it doesn’t result in duplicate records. So you can directly load the data from Data source to Info Cube.
Safety Interval Lower Limit: In same situation instead of upper limit, if lower limit is set to 10, then selections set by the system is 990 to 1100 (system will subtract from last delta numeric pointer).
If the delta field is Time stamp instead of Numeric Pointer, we can set the safety intervals in seconds (say 300s or 1800s etc..).
Note: If you use lower limit, it results in duplicate records. So you should not load this data directly from Data source to Info Cube. It is mandatory to use DSO in your data flow and then load the data to Info Cube.