DMR or Dynamic Cubes? Can Transformer or Cube Designer be used?

[gosoccer]

Hi Everyone, Sorry, I have multiple questions:

:) :)

We are about start into a new project and curious if using the Dynamic Cube implementation or
using the regular Framework Manager/DMR Model will be more useful?

1) We are also curious about the tool to use. I think we have licenses for Transformer and Cube Designer.

We have some Hierarchical Data presentation that we need to perform as a part of our reporting using
Report Author. For example: Top of the Tree: State then County, then City, then Town. We are showing
# of population and other information for each of these members.

State
       County
                 City
                      Town

2) If we start the regular OLAP environment using the Transformer and Static Cubes, is the establishment of the
overnight jobs or every four hours to run to refresh the cube possible using the Transformer in a Linux box or
is it very involved?

Thank you in advance for your time. :) :)

[MFGF]

Hi Everyone, Sorry, I have multiple questions:

:) :)

We are about start into a new project and curious if using the Dynamic Cube implementation or
using the regular Framework Manager/DMR Model will be more useful?

1) We are also curious about the tool to use. I think we have licenses for Transformer and Cube Designer.

We have some Hierarchical Data presentation that we need to perform as a part of our reporting using
Report Author. For example: Top of the Tree: State then County, then City, then Town. We are showing
# of population and other information for each of these members.

State
       County
                 City
                      Town

2) If we start the regular OLAP environment using the Transformer and Static Cubes, is the establishment of the
overnight jobs or every four hours to run to refresh the cube possible using the Transformer in a Linux box or
is it very involved?

Thank you in advance for your time. :) :)

Hi,

The choice of cube technology you choose really depends on the data volumes, the latency of the data, and the structure of the underlying data. The choices you have that are included with Cognos BI are Dynamic Cubes, PowerCubes and DMR. Each has a sweet spot and a use case. Here are some examples:

PowerCubes: Your data is held in multiple databases or files and data volumes are low. You need to keep the reporting load away from the underlying sources by extracting the data and building it into detached cubes. Build times are not a problem for you. Staleness of data is not a problem for you.
DMR: Your data is held in one or more databases and data volumes are low. You need data that is as fresh as the underlying database tables when reporting / analysing. Complex queries hitting the database(s) are not going to cause you performance issues.
Dynamic Cubes: Your data is held in star or snowflake schemas in a single data warehouse. Data volumes might be extremely high. You need excellent response times. You are able to add aggregate tables to your data warehouse as required.

These are not hard-and-fast rules you must obey, but they give you an idea of the best use-cases for each technology.

What do you mean by "as a part of our reporting using Report Author". Wouldn't every report be created by a report author? Or do you mean Report Studio?? ;)

MF.

[gosoccer]

Sorry, Report Studio. Thank you for the explanations. Do you know by any chance if the Date Dimension is created from an existing table that has all the date information. Like in Transformer, the dates where automatically created as a part of the cube for 1-whatever years I choose using a specific DATE COLUMN. So, the cube Time Dimension was created using the Transformer. 

In DMR, can the dates still be created the same way to automatically create the dates (I guess using the Framework manager) or the date Dimension has to be created in a different way?

Thank you again for your time?.

[MFGF]

Sorry, Report Studio. Thank you for the explanations. Do you know by any chance if the Date Dimension is created from an existing table that has all the date information. Like in Transformer, the dates where automatically created as a part of the cube for 1-whatever years I choose using a specific DATE COLUMN. So, the cube Time Dimension was created using the Transformer. 

In DMR, can the dates still be created the same way to automatically create the dates (I guess using the Framework manager) or the date Dimension has to be created in a different way?

Thank you again for your time?.

Hi,

Transformer, Framework Manager and Cube Designer all work in different ways. Transformer has a nice feature that will auto-create a time dimension (to the Month level) from a date column, and add relative time. It can do this because it is transferring the data to a physically different location (a cube file) so it's easy for it to restructure the data during this process. FM does not have this facility - all levels need to be specified manually in FM (often via calculations), and if you want relative time you need to code some more rather complex calculations. Remember that FM is pointing to existing data in its original location and reporting off this. There is no option for it to physically restructure the data. Cube Designer is about half way between the two - there is no "auto" time dimension option - you need to have a time dimension already present in your data warehouse, and point Cube Designer at it. You need to identify each level's approach (ie year / quarter / month etc), but you can get Cube Designer to automatically include relative time members and you can add your own custom relative time members too. These are added in memory in the member cache when the cube starts.

Cheers!

MF.

[gosoccer]

Thank you so much!
My questions have been answered.

Since the client doesn't want to deal with cubes, and we are not dealing with
a large amount of data, it looks like we will go after the DMR approach.

Thanks again. :) :) :)