| Replace, Re-Engineer or
Renovate This is not an easy choice to
make. Each answer has its constituency. There are legions of consultants ready to replace
"that old junk" with the latest and greatest, be it off the shelf packages,
semi-customized systems like SAP, or new implementations using the latest Java based
methods. There are re-engineering and renovation consultants arguing for "keeping the
best of the old" while providing varying degrees of a new look and feel. This might
be through some sort of middleware on the mainframe allowing browser access to CICS
applications, or by building a Java system on an application server which includes access
to selected mainframe data stores as well as its own database. Internally, there may be
groups who have invested a large part of their careers in building the existing
applications and who want to see them "fixed, not junked". The only general rule
that one can rely on for this decision is that unbiased advice is devilishly hard to come
by.
Regardless of the strategy ultimately chosen, the decision
should proceed primarily from business criteria: cost, benefit and risk. Each
representative for each solution will readily discuss the benefits, and will eventually
get around to discussing cost, albeit with varying degrees of accuracy. However, opening a
discussion of risk, both technical risk and business risk, may not elicit the full
picture. Indeed, IT has a less than stellar history of fully considering business risks as
well as performing overly optimistic appraisals of technical risk. Each solution has its
own unique risk profile, which may or may not be proportional to either cost or benefit.
Therefore, my colleagues and I tend to emphasize risk analysis and risk management in
making this decision, as a counterweight to those who underemphasize or ignore the issue. |
