I can give many reasons why someone wouldn't want Linux. Now don't shoot me, but you need to know the concerns when you are in meetings.
1. A new operating system, such as Linux, requires a new skill set. You either have to train people or you have to buy people to support it. Unless you bring in highly trained people, the Linux applications and availability are going to suffer during the learning curve. 2. Linux in production is relatively new. It is immature. Still better then Microsoft products, but not likely to be as solid as VM/VSE/MVS. 3. Mainframes are expensive. Ok not really anymore. I still have a gut feeling that a mainframe resource unit costs about twice as much as the same resource unit on another platform. However, we operate at higher utilization levels, lower support costs(one copy to support then a copy on each box) and less software costs (one copy per engine vs one copy each for each boxes) then most other platforms. 4. Linux isn't ready for prime time. Jury is still out on that. Jiffies patch is available to reduce overhead. We have a work around for the Linux memory management (we need a switch in Linux to turn off its caching) and we have some sections that can use VM's shared segments. Just look at this list for all the fixes and bypasses that seem to come out weekly. I'm not sure if the I/O performance problem still is around, or that it has been addressed. 5. They got burnt by going to client/server or Microsoft. 6. Support costs seem to be high. The way Linux is maintained might be ok for PC stuff, but I sure want a MSHP, VMSES/E or SMP to do this. I wonder what the Auditors think about this type of "shoot from the hip" maintenance? 7. Where there are many qualified people out there that know Linux, there are a lot more dam idiots that don't know a dam thing (I'm in this catagory, but I'm learning.) They may know how to put Linux on a PC and get an application running and talk about how this is the greatest thing and everything should be done this way, but the idiots fall apart when questioned about support, costs, performance, long term care and feeding, etc. and this makes Linux look bad to management. These are the things I'm trying to address plus to manage "managements" expectations. If they think everything is great with Linux and it doesn't pan out, you tarnished Linux in that shop. I would have much rather brought Linux/390 in more slowly on our new H30 then have our CIO start considering Websphere on the system (with the application licensed for 350 users...I.E. our GEAC accounting system). The first meeting on this was last Tuesday. I don't know much else about it. I'm trying to heavly suggest that I be invited to any meetings involving applications on Linux, and more so, Linux/390. Tom Duerbusch THD Consulting
