Hello,

We are at a very early stage of assessing the suitability of Baan Decision Manager 2.2.

At the moment we are looking at the basic requirements for this application. Does anyone know if there is a version of Baans Decision Manager that runs on UNIX ?

Thanks

Decision Manager does not run on UNIX system, but of course it works with Baan running on a Unix system.

Decision Manager makes use of a data warehouse that must be a SQL Server database. On top of SQL Server , Microsoft Analysis Services must be installed.
The data gets extracted from Baan in text files. Extraction is done by both sessions and exchange schemes. These textfiles are picked up by DTS in SQL Server to transform the data and load the Data warehouse. Since we are talking textfiles it doesn't matter on what platform Baan is running or which database Baan is using. DM2.2 is a set of tools (that runs on windows) to control this process. One of the tools, the modeler, is a development environment. Within the modeler you can generate the DTS packages, Baan Exchange Scheme's, etc. Baan not only delivers the infra as described above but also content. For each functional area, like sales, there is a template. Templates are sold sepperatly. These templates make life a lot easier. A data warehouse can be up and running in a day.
The front end tool to analyse your data warehouse can be any OLAP tool but Baan uses Crystal Analysis 8.5. See also www.crystaldecisions.nl.

Hi Jomanda

Do you know of any alternatives to Crystal that are possible or that someone may have installed

Hi Frank,

Baan used to have their own front end tool, Baan OLAP Navigator. Baan has decided to stop support on this product. That's why Baan was looking for a new (commercial) front end tool. They have chosen Crystal Analysis but DM22 will work with any front end tool that can connect to Microsoft Analysis Services. The cubes of Baan's datawarehouse solution are stored in Microsoft Analysis Services which means that any tool that can connect to Microsoft Analysis Services will work. Examples are Cognos and Business Objects. But keep in mind that these companies only deliver tools to analyze the cubes. Baan also delivers the templates (export sessions, exchange schemes, cube definitions, data transformations) to fill the datawarehouse and analyze the data.
I don't know of any customer that uses another front end than Crystal Analysis.

Success!

Kind regards,
Jomanda

Actually to build a datawarehouse and a cube is not so difficult, you do not really need to go with the expensive front end tools as Crystal if you do not want to:

e.g. for BaaN sales:
- export data from your tdsls051 tables to text files.
- import them into a table in SQL Server / Analysis Server (is easy)
- build a cube (is easy)
- use Microsoft Excel 2000 or later to browse the data.

Robert

Hi Robert,

What you describe will work but.......there is more to a data warehouse than copying a Baan table. For example if the Baan table is big you will want a mechanism for incremental uploads. Also how do you deal with multi site situations? This kind of logic is part of the Baan extraction sessions.

Another example is slowly chaning dimensions. For example, if the phone number of a customer changes you will probably want to overwrite the old number in you data warehouse. But....if the address of a customer changes you may not want to overwrite the customer record in your data warehouse because you may want to analyse per region in time. In this case an extra customer record has to be created in the data warehouse. This is in data warehouse techniques referred to as slowly changing dimensions. The customer table will need another primary key, a surrogate key which will be automatically updated during the upload.

You're solution will probably work fine for reporting purposes but if you realy wat to analyze there is more to it. To answer WHY questions many Baan tables are involved. For example if you want to analyze on the Perfect Order Rate, you will deal with Customers, Suppliers, Inventory, Shopfloor, Invoicing, Payments, Projects, Items, Time, etc. All logic to let these interact should be part of the transformation logic to upload the data warehouse.

All logic described above is part of the Baan Templates. Of course you will have to pay for them. It is up to the customer what they need, your solution, which is cheaper, or the Baan solution.

Regards,
Jomanda

Jomanda,

Again, things are not so difficult. Of course you will have to deal with these kind of issues, I merely gave a 'how-to-start' description.
Incremental uploads are easy using trdt fields as a selection for daily uploads, multi-site: we have a DWH consolidating 6 locations/ERP systems worldwide (6 currencies).

From what I've seen the BaaN Templates still very much mirror the design of the BaaN system, e.g.: there are seperate cubes for Finance transactions and for Integration transactions whereas you would expect 1 cube with a dimension which breaks down into 'origin': Finance or Integration. This is most likely because there is no direct link between the two tables in BaaN (IV).
The design of the DWH very much depends on what questions you actually want to answer, this will be business and customer specific. Big chances that you will still have to amend much of the design after a out-of-the-box installation, the idea of a 1 day ready installation is -IMO- a dream (and I'm not even talking about preparing the data export, making sure the quality of the data is good enough, making sure you have comparable data in case of multi-site situations; this can take as much as 75% of a DWH project).

regards,
R.