This has been the subject of several long message threads on the LinkedIn Mainframe Experts interest group. One discussion was initiated by Michael Healey around software by Mantissa. Another was initiated by Dave Austin and focused on SELCOPY and the likelihood of putting a Windows interface on a 3270 screen . These are two separate discussions but they all come back to Windows on the mainframe in one form or another.
Austin’s discussion was wrapped up pretty quickly. This is not a particularly difficult problem, and there are a handful of tools that approximate this to a greater or lesser extent. These, according to the various discussion contributors, include SELCOPY, ISPF, Bluezone, RUMBA, and various Free 3270 emulators. What you end up with is some sort of Windows-like display for a 3270 terminal. This may benefit folks who have to work through 3270 screens, but it is not a game changer in any way.
The game changer would be the ability to run virtualized Windows on the System z, much like you can run Linux or OpenSolaris today. I have asked IBM about that possibility and the reply has been that there remain technical and business issues to work out, most of them having to do with Microsoft.
However, that doesn’t discourage others from attempting it. The most interesting seems to be Mantissa, which is promising z/VOS . z/VOS is tantalizing. Mantissa suggests you can create a virtual PC in 15 seconds and have it operational in 15 minutes. According to the company, you could virtualize and manage thousands of PCs on a single System z with full Windows PC functionality.
There was a rumor at SHARE in Austin earlier this spring that Mantissa would demo the product. That didn’t happen. Mantissa has a solid track record of delivering mainframe products so I expect they can do it technically as, I’m sure, can IBM. The real obstacle will be Microsoft and what Microsoft’s army of lawyers can do to either prevent this from happening or extract a sufficiently high price to make it not worthwhile.
Of more concern to LinkedIn participants was the question of what you might do if you had the ability to legally run a virtual copy of Windows on a System z at a reasonable cost. This would be big indeed, as big as running Linux on the System z and certainly bigger than running OpenSolaris on the z.
Small x86-based server shops are not going to suddenly flock to the System z to run small PC applications. However, there are within large mainframe-based enterprises many pockets of stealth Windows applications. These run on small servers, often within workgroups and just as often below the radar of the IT group. The applications may not have Linux counterparts.
If the IT group could run virtual Windows servers on the System z they could consolidate these applications. Everyone would benefit from the reliability, recoverability, security, and manageability that are inherent in the System z. And the cost savings would be considerable although many of the costs probably are buried in hidden budget nooks and crannies throughout the enterprise.
Even when it is possible to run Windows on the System z, IT will have a tough challenge to reel in these stealth PC applications. IT will have to use brute force to smoke out these applications and bribery to get user groups to turn them over. The technical, business, and legal challenges posed by Microsoft will pale in comparison.