I think if you're simply trying to educate someone familiar with distributed servers, I would ask something like this:
"Are servers free?" And (hopefully) they'll answer, "No." Next question: "So how would you measure their costs? If it's a multi-user server, as most servers are, how would you allocate their costs among users?" And they might give a long answer which may or may not be insightful. Then I would simply say (assuming they demonstrated some insight): "There's really nothing different here. One of the ways you can measure and allocate costs is by calculating the number of CPU-seconds that a user requires and calculating a cost per CPU-second. That cost might be a very small number, but it's never zero. CPU-seconds certainly aren't the only way to base costs or even the best way. But it's one way you can do it. The same thing is true for electricity. Your electricity consumption is measured in Kilowatt-hours. One Kilowatt-hour might be very cheap but, added over many thousands of users and over many years, that small number adds up to a very big number, enough to pay for huge electric power plants. The same is true of servers of any type: there are underlying costs and, for every platform, one user might consume a very small amount, but the cost is never zero." Try something like that and see where it goes. Everything I said earlier still applies, though. Continuing: "Now, if CPU-seconds always cost some money, doesn't it make sense to try to save CPU-seconds if possible so you can get more users for the same cost? You can keep the lights in your house turned on all day while you are away, using electricity, but it will cost more and provide no value. Doesn't it make sense to get energy-efficient lightbulbs to save money and to turn them off when you are not using them? That's what performance engineering is about, and good performance is important on every system. On systems which serve the largest number of users, it's can be even more important. A little bit better performance can mean some large cost savings. Users like faster systems, too, because they can get more work done -- it saves them money because they aren't wasting time." And at that point, hopefully, the lightbulb will go off. No pun intended. :-) Hope that helps. - - - - - Timothy Sipples IBM Consulting Enterprise Software Architect Specializing in Software Architectures Related to System z Based in Tokyo, Serving IBM Japan and IBM Asia-Pacific E-Mail: [EMAIL PROTECTED] ---------------------------------------------------------------------- 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
