A recent Gartner Group report has some interesting things to say about VM and its future. Due to copyright restrictions, I can't reproduce the article here. I'll just quote the first paragraph and paraphrase some of the high points:
Everyone knows that VM is an unprofitable, dying operating system, that PR/SM and other hardware-partitioning schemes have ravaged the VM license base, that VM-based office systems are terribly expensive compared with newer LAN technologies, that VM has no significant growth potential, that VM code quality problems are pervasive, and that most customers are leaving VM. Curiously, everyone is quite wrong.
See Gartner report: 93/04 GG-SMS Six VM Myths Exploded, SMS Products, P-507-1291, April 5, 1993, W. Malik
You also might find the this report interesting as well: 93/04 GG-SMS Facing the Client/Server Locomotive, SMS Key Issues, K-557-1288, April 5, 1993, A. Percy
Summary:
Clients committing to move from mainframes to client/server (a false distinction) need to reinspect why they had mainframes in the first place.
Gartner Group, I must mention, occasionally reverses themselves on this. They interpreted the moves of some mainframe clone competitors, plus IBM's not providing SYSPLEX for VM as implying IBM was dropping support for VM. IBM's Endicott Labs then had to go explain this all to Gartner; including the idea that perhaps VM needs SYSPLEX as much as a fish needs a bicycle. I admit I take Gartner Group pronouncements with a grain of salt.