> I think that you are talking about something that is either going to hit > us real hard or IBM is going to come out with something that will > eliminate the need to the CMS based tools "old folks" such as me and, > having met a lot of you at SHARE conferences, most of the rest of you.
One of the questions I asked George Madl in his VM Directions session at zExpo in Munich was that given that there seems to be no further roadmap for additional CMS investment, what the migration plan might look like for CMS users to (probably) a Linux environment as the "personal operating system" for interactive users. He didn't have an answer, but asked me to assemble a list of things we think we might need. I think it's pretty clear that we will need ways to build and maintain CP from Linux (the TPF guys have a pretty good head start on this one), and we'll need some REXX and CMS command utility emulation to provide a moderately smooth migration for our execs. We'll need a formalization of Linux access to CP services and capabilities, either by a common API or by REXX and Perl function packages. We'll need at least emulation of the linemode capabilities of XEDIT (a full-screen emulation that is termcap-aware would be awesome, but a lot harder), and some kind of emulation for CMS Pipelines. I think we'll also need tools to migrate compiled modules -- sort of a Cygwin for CMS applications; intercept the CMS APIs and emulate them. Other ideas? I'd be very interested to know what others think about this. > Maybe this means that IBM is going to eliminate the need for us CP/CMS > knowledgeable sysprogs. One of the basic value points of the combination of LPAR and VM is the ability to virtualize resources at both a macro (LPAR) and micro (virtual machine) level, which is much finer control than is present in any other virtualization solution. I'd expect more a plan to finally make VM a ubiquitous feature of the hardware -- at the current price points, and given the withdrawal of VSAM pretty much kills CMS as an application support and testing platform, layering the cost of VM development into the price of hardware doesn't seem to hurt much and it's a huge PR win vs VMWare or Xen. Heady stuff. -- db
