Datasheet

Chapter 1: SSIS Solution Architecture
15
Estimate your data volumes in one of the initial development cycles so that you can purchase
the right hardware.
Get your server storage ironed out upfront. Be sure to set expectations with the storage group or
vendor early on in the process.
Plan out your package storage and deployment plan in one of the initial phases. (Chapter 3
provides an in - depth discussion of this topic).
In every development cycle, be sure to include a data - validation task so that you can have data
checkpoints along the way, rather than having one data - validation test at the end (which often
leads to changes).
In regard to SSIS data - related solutions, you must plan upfront any initial data load
requirements. If you leave out this planning step, you will likely underestimate the overall
solution scope.
Data Element Documentation
Not many developers or system architects are fans of documentation or at least writing
documentation. However, it is a necessary task in any data - centric or ETL project.
Again, this book is more about SSIS solutions than project management, but given the importance of
tracking data, included here are some recommendations on data - tracking documentation that can help
you in your project.
Data documentation is about tracking the source and destination data elements, data profiling, and
mapping the data elements to requirements. You must be diligent about these tasks, and keep them up -
to - date, because doing so can help you keep control of the data your project uses. Documentation is also
useful in future administration and lineage.
The following data - tracking documents should be used above and beyond your usual documentation
(requirements, conceptual design, physical design, ETL design, and so on).
Source Data Dictionary, Profile, and Usage
The source data dictionary is about more than just a definition of what the data elements represent. It ’ s
also about what the data looks like, and what destination elements it is used in. Planning sessions can
then refer to the source dictionary to help validate requirements and data availability.
You should structure this in two sections: entity definitions and element definitions.
Table 1 - 1 provides some details for entity tracking.
c01.indd 15c01.indd 15 9/24/09 11:26:35 AM9/24/09 11:26:35 AM