Quote - Assuming an additional administration overhead of only 20 minutes a week for a Win PC makes in 20 years about 340 hours. I charge 200-400 bucks for my business hour and I'm my own sysadmin. The bottom line is: I saved around 100.000 $ because I've been using a Mac instead of a Win PC in the past 20 years. If I consider the superior hardware quality and backward compatibility of my Mac, the calculation gets even worse for the Win PC.

You oughta see it from my end. (warning: ungodly geek soapboxing ahead...) An average MCSE can feed and care for about 50-60 servers at a go, maximum, before their schedule gets choked up w/ patches, A/V updates, compatibility testing before each update, closing security holes post-update, and having to deal with every little breakdown that occurs. OTOH, I could readily care for and maintain the 150+ RHEL Linux, FreeBSD, and Solaris (and one Windows) boxes at my last employer, no sweat. The biggest part of my day usually consisted of a couple hours checking log results, firewall stuff, backup reports, and looking into any SMTP (email server) bounces to insure that any hangups get cleared. I have never seen any security-related settings open up because of a patch in *nix or Linux. The vast majority of my day could be spent listening to my users, tweaking the system when needs changed, and looking into new ways of getting something done. Downtime? The only time a *nix box needs rebooting is if a) the hardware fails, or b) a kernel patch goes in... and the downtime for that can be scheduled at sometime in the future about 999 times out of 1000. Windows Server 2k3 needs a reboot roughly once every 3rd patch (sometimes more, sometimes less). Each reboot (on either server) will eat about 10 minutes of downtime at the very least, assuming everything goes well. If not, you simply roll back to a previous kernel patch-level come next reboot in *nix to make it available for use while troubleshooting, but in Windows you need either Safe Mode (if you're lucky), or the Recovery Console and a lot of troubleshooting while the server is basically down (if you're not). Time spent on A/V? Well, there are a/v scanners that work in Linux (uvscan) that are sometimes required by the powers-that-be, but that can be scripted and basically ignored unless something pops up in the security newswires that needs a closer look (somewhat rare, but it happens). There's also an on-the-fly A/V scanner for mail services (such as ClamAV) which can be scripted to update itself so as to protect your Windows users from email-bourne crap... so I guess that counts a little. Security I could also compartmentalize security easier at the server level by editing one config file (iptables) and pushing it out, than the average MCSE could by clicking-in rulesets one at a time in IPSec, one server at a time (which many wouldn't even know how to do in my experience, so they rely on external products such as Checkpoint). There's also licensing... Windows Servers have these nasty little things known as CALs (Client Access Licenses), which can really add up as you pile on the servers. This is prolly why MCSEs (of all skill levels) average about $40k while the *nix admins (ditto) average $70K+ in salary... we can do 3x the work in less time, and don't require as much hardware muscle to get the same jobs done. Basically, you only need 1/3 the staff and 4/5ths the hardware budget to do the same job if the staff is competent. (whew...okay, all done now :) ) /P