I challenge the premise of reducing costs by movi9ng from mainframe to other platforms. We just did a study on Websphere TCO and concluded that the mainframe would cost less and the service quality and performance would improve substantially. We are now studying z/LINUX to replace Windows and UNIX servers, especially in light of our “GREEN” initiative. Our facilities analysis shows that mainframe will yield a 20x (YES, twenty) improvement over windows servers, and a 10x improvement over UNIX .
However, we have the advantage of already having multiple mainframes, so adding the new speciality engines is quite cost effective.
We all know that IT departments can spin a business case any which way they want. If you do a really thorough anaylsis of ALL costs, I believe the mainframe would surprise many CIOs. Get IBM to help with a TCO study. Make sure you include an analysis of downtime which requires a valid user determination of the cost of being down (productivity, loss of business, etc.).
Been involved in an MVS to Client Server migration. But the cost associated with the software will change little. Most licensing will cost the same regardlesss of the platform. Your savings will be in hardware and whatever other applications you can migrate to non-mainframe type of equipment.