Bill,
> -----Original Message----- > From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On > Behalf Of William Ball > Sent: Thursday, 14 July 2005 6:24 PM > To: [email protected] > Subject: Re: Another - Another One Bites the Dust > > You're jumping to the conclusion that squatty boxes are -cheaper- than a > mainframe, while in fact they are -not-....overall. They may initially > cost less than a mainframe but by the time you add a 5+ to 1 people cost > per application to the squatty end, they are actually more expensive. I'm jumping to conclusions - people in glass houses... Where is the evidence that there is a 5:1 ratio of people overhead for Unix and Windows. It may be your site's experience for one reason or another, and I see lots of hearsay on this list about it. But I am in and out of large Unix sites in Asia all the time, including converts, and my experience is that this is just not true. In fact, as skills and software improve in Unix I am starting to see the opposite - people bloat is on the MVS side. Look at your own example - you Unix/Wintel environment is at least 8 times the size of MVS, but there are only 5 times as many people. That sounds like a promising number for Unix doesn't it? > > And the "smoke and mirrors" that squatty box people pass off to the end > user, becomes a nightmare for the end user. Management by magazine has > done a great job of convincing these people life couldn't be better.....if > they'd just get rid of the mainframe. > What smoke and mirrors are you talking about specifically? You are telling me that you can tell the difference when a POS machine or ATM connects to a MVS, Sratus, Unix or Wintel - get real! Would Google or E-Bay users be happier if it was running on MVS? I think not. > Take a guess. If a current report, on the mainframe, takes 24 hours to > print, how long will it take to print it on a printer in the user's > department. Come on, you really think that the platform that a printer connects to will change the speed at which it will print? Connecting a 3900 printer to Windows, AIX or MVS is not going to change the IPM. > > The squatty nightly backups run (when they actually run) 12 hours and they > are just adding another 8 TB of disk. The mainframe has 1.x TB and the > applications going on the squatty boxes are -replacing- ones already > running on the mainframe. You can't exactly blame the squatty boxes for poor planning. Up to recently there was an MVS site with a backup job that started at 7:00 am, and it was usually still running when the applications batch jobs started the following night. You know the typical XYZ.** backup to one tape and TOL(ENQF). Your point would be that all the Unix/Wintel staff at this site should expect all MVS backups at every site to be just as inefficient (this was just 2TB). Personally I think that Unix and Windows backup software like Commvault and Veritas leaves MVS backup software looking like something from before the wheel invented. Three year ago I was part of setting up a video backup system for a casino that had to sustain 1GB/sec (1000MB) backup throughput for 5 minutes out of every 15, 24x7. I don't think MVS was ever going to be the outstanding TCO winner for that environment. > > What do you think is it going to be more or less expensive in the end with > a squatty solution? > The argument that TCO for most applications would be less on MVS than Unix and Windows seems to lack one basic ingredient - the meat! Companies are not voting with their feet because someone read a magazine on a plane. This management by magazine myth is one incredible piece of crap, and not worthy of this list. Some of the incredible strawmen (not you Bill) pitched up in IBMMAIN suggests that the total research on TCO for Unix systems by MVS System Programmers has not gone much further than the IBMMAIN archives. Ron ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
