Hi Sarvanan,
At present i am learning by trial and error, and with actual hands on experience of On-Demand Solution:
Observations so far:
1. On-Demand solution is still evolving from its original form of - on site or CD solution, so Data Dictionary is good as a starting point, however, it is not a reliable document . There are factual inconsistencies between what is required to make a successful file load, and what is available in data dictionary.
2. Ariba provided standard extraction program ZARIBA_SSP_MASTER_DATA_PULL fetches only about 40% of the total files you would need to make On-Demand solution to work. Balance 60% files, you have to prepare manually. AND IT IS CUMBERSOME!!!!
3. Information which is being pulled thru the extractor program are not necessarily in sync with the data dictionary column structure, which means before loading the files, the project team needs to either add or delete the columns.
4. when you log in into Ariba the drop-down menu options has weird export parameters for the field values. For example in GL account, you can see only the GL account numbers ... No description, .. Imagine u viewing jus numeric value and being asked to select the a G/L account for raising the Requisitions.
5. If you have worked with SAP and SRM, some of the points like 'delivery address' and how to enter delivery address, are significantly different in Ariba. In Ariba, a 'plant' is created to represent a delivery address. ( Imagine you want to allow user to enter ANY delivery address and you would end up creating sizable number of plants.
Hope this would give you a flavour of challenges your project would get while implementing Ariba On-Demand solution.
If you have any specific questions, please let me know .
Regards
Rahul