Extended Star Schema doubt

Question

Hi ALL,
Fact table --- Dimension table -
Sid Table -- master data values.
Fact table:
Contains Key Figures and Dimensional id's
Dimension table:
Contains Dim Id & SID
SID table:
SID and Master object ID ( eg: Customer ID)
Master Data Table:
contains Customer Attributes , Texts , Hierarchies for that Customer ID.
THis is the Extended Star Schema design.
MY DOUBT:
1)In Dimension table itself if we place that Customer ID(No SID table next) what will happen?
Fact table --> Dimension table --> Master data Table
2)Instead of that SID table we can directly place that CustomerID in Dimension table , so we can reduce one layer inbetween Dimendion table and Master data table.Is it correct or not?
Any one can clarify my doubt.
Regards,
Arun.M.D

Answer 1

SID means 'surrogate ID'. That is an system created id as you know. Main purpose is fastening the search.
Mostly, there exists a rule for Customer or Material ID's.
Like it should be CHAR 10 or CHAR 16.
This kind of alpha-numeric fields are harder to search when compared to integers. Moreover, your customer id can be 10 digits but, this does not mean you will have 1000000000 customers. This is the main reason that, an internal ID is produced. If you have 10000 customers, your SID will be at most 10000. 
However, if your customer ID's are starting from 1 and growing up like integers, then your argument would be true. ( but still no way to skip SID creation and direct usage of characteristic ID in fact table)
Also, as mentioned by other friends, there exist the Line Item dimension property if you have only one characteristic in one dimension. That simply does skip the DIM ID creation step, and puts your SID into the Fact table. ( Since you have only one char in the dimension, no combination is possible)
Hope this helps.
Derya