When a new ERP implementation is planned, it is usually a well planned, multistage process. Based on my experience of converting legacy data into Oracle E-Business Suite, I am going to share a few things about this process.
By the way, I have spent about 7 years of my life converting legacy data into Oracle E-Business Suite for almost all modules of Oracle E-Business suite. I have worked with teams in US and many international countries.
I am not going to discuss general steps of planning and various standard phases of the implementation cycle because those are standard and you can’t avoid them. However, I am going to discuss a few technical aspects that can make the implementation, a repeatable and robust process.
Let’s say an organization is starting a vanilla implementation of Oracle E-Business Suite. You plan your CRP phases, you plan all customizations and everything.
Now, when it comes to extracting data from legacy systems (and situation gets tricky when you have multiple legacy systems to convert), how do you model the solution?
In this article, I am going to discuss a few steps that I have found to be very effective in making the whole data conversion process, very repeatable and cost effective.
Steps To Design An Effective Data Conversion Process
1. What Data Needs Conversion?
The first and foremost question to answer is, what data do you need to migrate? Which Oracle ERP modules are to be used by your organization?
The answer to this question depends on the nature of business and the nature of ERP implementation. But, it is very important piece of information to identify before designing the data conversion strategy.
Once you the data to be converted is identified, the next step would be to design the overall data conversion strategy and to design the solution even before setting up Oracle ERP system. This allows the organization to optimize resource usage.
2. Design The Data Conversion Solution
This might sound easy because the basic understanding is that Oracle E-Business Suite comes with a set of open interfaces and APIs to convert data into ERP once data is placed into those interface table.
However, the number of validation the legacy data has to go through before it could be really converted, is huge. This is where the complexity comes. If you don’t plan and design a robust solution, the data conversion itself can become a very resource intensive and costly affair.
So, I recommend (based on experience) that you follow these high level steps to make this process simple to understand and repeatable (especially if you have multiple company sites to convert):
1. Define a common format (such as a flat file format or XML format) in which data from legacy systems will be extracted for conversion. The data format should be comprehensive to cover each data field for every module.
2. Design an ETL engine that can read data in above defined format and place it at a central location where validations could be run to ensure data quality. This allows you to extract data from source systems and start running basic data quality validations (such as NOT NULL fields, Field Length checks etc.) way before even the ERP implementation teams are deployed at the site.
3. Extend the ETL engine to take cleaned data from the central location and validate against ERP setups, transform source data into a format that Oracle ERP open interfaces would accept and finally, load valid data into the Oracle open interface tables.
4. Have Oracle ERP setups completed and concurrent requests ready to run the open interface programs after data is loaded into the interface tables.
5. For interfaces that do not have ready to run concurrent request programs (or the modules that are known to have problems), create concurrent programs using Oracle APIs to load such data elements from the open interface tables or custom tables as your situation demands.
3. Deploy The Data Conversion Teams
Once the data conversion engine is ready and processes are documented, it’s time to deploy the data conversion engineers to the site which needs to be converted.
The role of the team is to extract data from the legacy systems in the format specified by the conversion engine specifications and run the load and validation routines repeatedly until legacy data qualifies to the minimum quality standards needed by data conversion validation routines.
Once basic validation is done and data follows minimum quality standards, it’s time to move to the next step of the data conversion ETL which will validate the legacy data against Oracle ERP setups for your organization. After this step is correctly completed, data is pushed automatically into the open interface tables for final conversion by standard/custom routines.
4. Final QA & Production Loads
Once you have completed a couple of rounds of data conversion with the legacy data depending on your overall project plan for the site in question, it’s time to take this process into QA environment.
The data conversion in QA environment is supposed to be as if it is happening in production. Data that has been already corrected during earlier phases is passed through the data conversion engine ETL against Oracle ERP production setups.
Once QA loads are successfully completed, you are ready to take your legacy system live into oracle ERP system.