Q:  

What are the disadvantages of having a data warehouse in place? Is the disadvantage connected to the data in the source system with regard to the cleanliness of data or any other issues involved therein?

A:  

Chuck Kelley’s Answer: I am assuming you are asking about using the "virtual" data warehouse approach. This approach says you do not need a separate data structure, just some middleware that understands the data. The issues I think are:

  1. Do you really want to allow users to issue queries into your transaction system that will do sequential scans of the data? I think not.
  2. Do you really want a middleware program trying to figure out the different meanings of data (Gender = M/F, or 1,2 or 0,1) and, at run time, peruse the transaction systems to find it?
  3. How do you tune a database and the operating system to do three reads and five writes (typical transaction) AND read a million rows and aggregate at the same time? How would you index it, for example?
  4. Cleanliness of the data is an issue, as you pointed out.
  5. What if the keys to the major data (customer, product) are different? How does the middleware propose to deal with those issues?

These are just some of the issues concerning your question. While the concept is interesting, I do not believe the technology has caught up with concept.
Larissa Moss’s Answer: It is not possible to talk about "general disadvantages" of having a data warehouse. Let’s look at why a company would want to build a data warehouse in the first place. Presumably the company has information needs that are currently not being satisfied by their operational and decision support systems. Presumably, the company is suffering from substantial losses due to the non-availability of this crucial information. These losses can be anything from losing business to the competition, losing customers to cost overruns, etc. Presumably, the reason for not being able to satisfy their information needs is due to dirty data, redundant data, inaccessible data, untimely data, etc. In other words, the company suffers real business pain, which it cannot solve with its current systems. This pain usually translates into high losses. Once the company evaluates the cost of a data warehouse solution (which addresses all the data issues when done correctly), the company has to decide if the benefits (profits to be gained or losses to be eliminated by having their information needs satisfied) outweigh the costs of building a data warehouse. If it doesn’t, a data warehouse solution is not appropriate and maybe it shouldn’t be built. If it is built anyway, I suppose that could be considered a "disadvantage" since it has no payback.

Clay Rehm’s Answer: The only disadvantage of a data warehouse is one that is poorly designed! Indeed, a data warehouse will certainly introduce data quality issues as well as many other processes, data and resource changes – however, all for the better. It is important to understand that the maintenance of a data warehouse can be a time-consuming and difficult task. A company must make the commitment to supply a dedicated support team, appropriate hardware and software, and business sponsorship to make it effective. Even though building and maintaining a successful data warehouse is a difficult job, the benefits far outweigh the expense.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access