Kimball Forum
Would you like to react to this message? Create an account in a few clicks or log in to continue.

Operational Reporting : DWH vs Legacy Application on Mainframe

2 posters

Go down

Operational Reporting : DWH vs Legacy Application on Mainframe Empty Operational Reporting : DWH vs Legacy Application on Mainframe

Post  pbestgen Thu Feb 05, 2009 9:26 am

I'm BI Project Manager in Organization and Process department in my company (ie from the business side and not IT side ).

We have implemented a DWH with dimensional modeling. It contains absolutely all the data handled by the legacy applications on mainframe. The data are loaded daily into the DWH via ETL procedures.

For several years users have created themselves their own operational listings/reports based on the DWH with BO Reporting tool. These listings are critical because they often allow to control and monitor the daily activity of the business. Often, the content is the lowest level of granularity of fact tables.

However, IT peoples believe that these operational listing should be produced by the legacy applications on the mainframe. I consider myself that end users in the business are able to design such list with a tool like BO and that is cheaper than asking a computer specialist to program it in cobol.

Does anyone could share his experience and give the arguments in favor of operational listing based on DWH rather than source applications.



Posts : 4
Join date : 2009-02-04

Back to top Go down

Operational Reporting : DWH vs Legacy Application on Mainframe Empty Re: Operational Reporting : DWH vs Legacy Application on Mainframe

Post  Edwin Kurian Thu Feb 05, 2009 11:55 am

First of all let me preface by saying that generating reports on the legacy systems (operational data) is like the prenatal stage in the TDWI BI maturity model. You may want to check with the BI folks in the IT team to understand what their thoughts are. This situation generally occurs if the BI team is either non-existent, does not have the requisite skills, has difficulty to influence decisions, or has a management team that does not have a clear understand of the benefits of BI.

Certain kinds of processing and requirements belong to one environment and other kinds of processing and requirements belong to another environment. As IS organizations mature, they progress from a mind set of having one computer environment serve all the information processing needs to the realization that multiple, specialized environments are needed. The first two environments they discover that are needed are the operational environment and the Data warehouse informational environment. Day to day decisions is made on operational data, medium to long term decisions are made on the data warehouse. Operational data contains detail data, whereas the data warehouse contains some detail and some summary (read query performance)
Operational data is architected 3NF to facilitate insert/update/delete. Data warehouse is mostly architect in a star schema to facilitate query reads. Query performance is a key criteria for the data warehouse to succeed.
Star schema is flexible and scalable as requirements are added. You can easily add more dimensions and attributes.
Operational list is accurate at the moment it is accessed. If a business condition changes, that list may not be accurate. On the other hand a data warehouse contains historical information. Using concepts such as slowly changing dimensions, reports will produce the same results consistently.
Operational data is rarely clean, consistent, or easy to integrate. It also has multiple versions of the truth. This data is cleaned when it gets to the data warehouse. Using conformed dimensions you have a single version of truth.
By running operational reports, you are sharing hardware/software resources with a transactional system. Having a separate environment eliminates (umm...mitigates) resource constraint.

There are lot of other reasons to use the data warehouse rather than the operational data. I would highly recommend reading the TDWI article "Smart Companies in the 21st Century: The secrets of Creating Successful Business Intelligence Solutions" By Wayne Eckerson.
Good luck.

who believes that the BI team needs a strong champion to succeed.

Edwin Kurian

Posts : 13
Join date : 2009-02-03
Location : Milwaukee

Back to top Go down

Back to top

- Similar topics

Permissions in this forum:
You cannot reply to topics in this forum