Problem solve Get help with specific problems with your technologies, process and projects.

Determining the right DBMS and OS for a data warehouse

I'm looking for a checklist that an an application area might use to help determine the most appropriate DBMS for either a data warehouse, data mart or OLTP application. Would there be a similar checklist for the most appropriate OS and hardware platform? How does NT or Windows 2000 Server measure up as an OS environment compared to a Unix for running Oracle or UDB or Sybase?

As you might expect, the main players in the data warehouse space are Oracle, IBM and TeraData. Each have their strengths and their challenges. From a checklist perspective; we recommend you take the following into consideration:

- Is the pricing by named user, concurrent user, by server, are site licenses offered, enterprise licenses.
- What is considered the DBMS ceiling with regard to database size? Is the DBMS really designed for OLTP performance (high transaction volumes against small numbers of rows during inserts, updates, deletes, reads) or data warehouse performance (high data volumes for bulk inserts, bulk reads). What utilities are available for bulk loading, options for loading data components separate from index components. What are the indexing schemes supported?
- If your data warehouse is starting out small and you anticipate it growing with control over a period of years, you might want to consider starting out with an NT solution running something like Oracle that can migrate to UNIX with very little changes to the schemas, structures and functionality. If you are a predominantly IBM shop, the UDB product suite provides that same level of platform portability all the way up to the big iron.
- What tools does the DBA have to monitor and tune? Cold spots, hot spots, index performance, optimization, capacity utilization, growth, security, access, roles, synonyms, etc.
- Does the DBMS come with an extended feature set targeted for data warehouse functionality? How proprietary is that extended set?
- Is your organization into very complex regression analysis, multi-dimensional analysis, statistical modeling or is it more of a medium level of complexity? You'll need to consider some of the multi-dimensional DBMS such as Redbrick or taking advantage of the "virtual" features offered by some of the BI tool vendors, such as Business Object's Universe for your data mart implementations.

Check list of things to consider when determining the appropriate OS and hardware platform. The following 6 steps can be applied in general when working with any server/disk configuration. The list helps weed through some of the manufacturers. Usually what is currently installed sets the standard for what will be acceptable in the future. But this shouldn't be assumed. Challenge your enterprise based on facts. These six steps will help support your case.

1. Create a list of server and disk manufacturers that are acceptable to your company's "stated" architecture and/or purchasing policies. At BAG, we encourage our clients to at least consider HP and SUN on the UNIX side and Dell or HP on the NT side. For UNIX storage consider EMC in addition to your selected server manufactures disk solutions. For NT you can consider some of Dell's network storage devices.

The response is continued...

Dig Deeper on Oracle data warehousing

Have a question for an expert?

Please add a title for your question

Get answers from a TechTarget expert on whatever's puzzling you.

You will be able to add details on the next page.

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.