>Also, I was nodding in agreement with John McKown up until he wrote that "most 
>successful migration projects [from mainframes] deliver a positive ROI

Actually, you blame the wrong John for this quote.  And it was somewhat taken 
out of context.  I was careful to point out that many migrations fail to 
deliver the quite rosy results initially claimed.  And also that many such 
projects fail, for the same old reasons many other IT projects fail.

But in my own experience,  the migrations on which I had some involvement that 
were technically successful, were also financially successful (and not by 
stretching it out for "centuries").

I have also been involved in two failures.  One failed because the prime 
contractor failed to perform due diligence and found themselves with a loosely 
defined scope and some migration issues for which there was no easy solution.  
The other failed due to deficiencies in the chosen tool regarding data 
integrity, and concerns about the migration tool vendor's ability to provide 
future support.

I have also heard of several others that were successful, both technically and 
financially.  First and foremost in my mind is the NYSE, which migrated a major 
trading application using Clerity (UNIKICKS) back in 2008.

IMO, the participants on this list are unnecessarily fearful of the risk that 
legacy migration poses to their livelihoods.  For these reasons:

(1) Migration of a legacy application is almost always Plan "B".  Or even Plan 
"C".  Most projects start as a desire to completely replace the aging apps with 
brand new Java or DotNet custom apps.  Or go with an ERP package like SAP or 
Oracle.  Only when they get over the sticker shock for these alternatives do 
management start thinking about the half-measure of a migration.
(2) Migration technology is mature only for the "sweet spot" of COBOL/CICS/DB2. 
 Start throwing in IDMS, ADABAS, NATURAL, PL/1,  ADSO, IMS-DC, or any of a 
thousand other things that are fairly common, and you quickly get to the 
"bleeding edge" of migration technology.
(3)  Sometimes migration is just another dead end, only cheaper than the 
original.
(4) It is hard to do migrations piecemeal.  So many are "big-bang" 
implementations, which are risky.  So the risk-adverse clients like banks and 
insurance won't go for it ever.
(5) Plan "A" for many companies is to merge with another.  One result of such 
mergers is that often there is a merging of IT technologies over time, possibly 
resulting in lights out for more mainframes.
(6) Successful migrations require people skilled in both the legacy and target 
environments.  These are rare.
(7) Many IBM listers fail to realize that Legacy Migration is a career 
opportunity for them.  So even if more clients start doing them, they can 
leverage their legacy knowledge by participating on such projects.

John Roberts

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to