Databases Reference
In-Depth Information
FOR EXAMPLE
Data in Transition
You might wonder about the justification for two- and three-tier data when
a centralized environment with a single database and a single server is obvi-
ously easier to manage and maintain. The decision to use a multitiered data
environment is usually one made more out of necessity than out of choice.
It's easy to think of the transition from PCs hosting their own data, data
stashed away in different applications such as spreadsheets, or from a tra-
ditional mainframe, to a PC-based DBMS as something almost surgical in
nature. One day, you're operating with all of these different sources and man-
ually compiling the results, and the next, everything is sitting in the data-
base server ready for use. The reality is often very different. The transition
is often more of a gradual migration than a sudden switch.
There are several issues that must be considered and real-world factors
that must be included when migrating to a new data environment. Even
though you are moving data from PCs to the database, the data is contin-
ually changing. It may be necessary to run different systems in parallel until
the changeover is complete. The transition could be made in phases,
putting the most critical data or data needed by the largest number of users
onto the database first and then gradually moving the rest of the data
as time and resources allow.
There are special issues that can arise when dealing with an existing
mainframe or minicomputer. A company's options are often limited by long-
term lease or service agreements. Maybe when you acquired another com-
pany and its assets, the mainframe contract came along as an unavoidable
part of the deal. You might not need the mainframe as your primary data
source after you've extracted the data from it, but you have to keep paying
for it, so you might as well get some use out of it. In other cases there are
practical reasons why you can't move the data to a PC-based database server.
The transition might be too expensive to make it cost effective, or the
mainframe data is a required source for a proprietary or third-party appli-
cation that runs on the mainframe only. It might be that communication
with customers or clients is mainframe-based and it is either too difficult
to move to a different method or there are licensing, patent, or copyright
issues that prevent it.
The reason really doesn't matter. The bottom line is that there are often
valid, unavoidable justifications for maintaining data in a two-tier or three-tier
configuration, even if just temporarily “during the migration.” It's important
to understand, however, that “during the migration” has become a permanent
way of life for some companies.
