- 26 -
operations on the same data. For the following scenario, let’s assume that our data is
stored in a common relational database management system, let’s say a SQL Server.
You may want to perform both complex statistical analysis and simulation analysis on
your data. Programs specifically devoted to each part are available (for instance, Minitab
and Arena), and each application is easy to use in its own right. The only problem is that
you don’t have a constant way of exchanging information between the two. Minitab
processes data in its own native format and Arena does the same. For argument’s sake,
let’s say that the results of one are dependent on the other (that is, they need to
exchange information between each other). One alternative (and this is what usually
happens) is for each program to do more than just one type of processing and to
essentially perform the same tasks as the other does, but that’s a really bad solution. It
results in a lot of redundant resource allocation as well as maintaining inherent
inefficiency; after all, each of these programs is specialized.
Another solution is for each vendor to create a translation program that allows it to
directly access information stored in your database. However, this is more difficult than it
sounds, and the vendor has no way of knowing up front that the information you have is
stored in Microsoft Access, Oracle, Sybase, or any other DBMS vendor. Figure 2.5
summarizes this dilemma.
