On Tue, 2004-04-13 at 14:10, Lionel Dyck wrote: > I was wondering what y'all see as a reasonable number of FTEs for 'n' Unix > servers and how that > changes with virtual servers under z/VM using Linux.
Experientially, being part of the system administration staff where the ratio is greater than 50 to 1 is no fun at all. 30-40 for reasonably well-behaved Unix servers with few or no end-users actually logged on seems about right. Half that for machines with significant end-user involvement (10-15 seems to be about right for having to do, say, Windows Desktop Support (ugh!)) that isn't restricted to, say, "the IMAP API". I tend to say the number probably doubles under z/VM, because management gets so much easier. But what that probably *really* means is that your VM Guy (with proper training, of course, plus a Linux guy if your Mainframe Guy is your z/VM and z/OS guy) can take care of any reasonable load on a single zSeries. Now, this would change a bit for something where you had a lot of machines but a light-duty usage cycle (like LCDS: lots of machines, most idle at any given instant, but heavy end-user shell activity). But even in that case, you probably won't have that many users with problems simultaneously. And in an environment like that, it's probably OK to tell your users "we are not going to solve your problem. We are going to reimage your machine instead." So, as a rule of thumb, YMMV, no warranty, etc.: 30-40 units per FTE for 'n' discrete-box Linux or Unix servers, probably 75-80 for a zPenguin colony. For workstations rather than servers, 10-20, depending on how adventurous your users are. Adam ---------------------------------------------------------------------- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
