Hi Salva,
The answer is, It depends! As has been said in other replies, the answer is, at the very least political. I used to work with a retired IBM capacity planner. There are two main purposes to a billing system. 1) recover your costs and 2) encourage sustainable computing behavior. 1) costs, all costs. This doesn't mean that you have to have a separate billable for every little thing that goes into cost. We have a CPU charge that incorporates support staff, equipment maintenance, operations staff, etc. We also charge for DASD excps. We used to charge for DASD residency 25+ years ago, but it caused undesirable user behavior so we stopped charging for DASD residency. There is still DASD cost of course, but it is included in the rest of the rates. It's also important to have rates that your customer can understand. We basically charge for CPU, DASD excps, tape, tape mounts, print records, and a few other things. When we used to have a lot of tape drives and operators to mount tape, we had a lot of user tapes to deal with. Now, almost all of our tape activity is virtual. Our billing charge for tape handling is much higher that the charge for FTP data pickup and delivery to the user's server. Print records sent to a viewer are less expensive than printed reports. Customers can also print all or part of their reports on their own printer, remotely. Remote print records are billed at a lower rate, but we still have to recover costs. Everybody has to recover costs somewhere, somehow. 2) encourage sustainable computing behavior. For example, if you charge for DASD residency, consider making the residency charge less for the cheaper disk, more expensive for the fastest, more expensive, etc. Another site announces when their oldest disk is coming to end of life - at that point the maint prices for it are generally going up - so they bill more for it and that helps motivate their customers to want to move to other disk. The price increases cover the extra cost and the natural desire of the customer is to avoid the extra cost and be willing to have their data moved. So if CPU is more expensive for batch work at peak times, customers will be encouraged to move what they can to other times - and that will help you keep your rolling average lower and save you money too. For other types of work, programmer testing, etc. the temptation is to run that workload at really low priority, but programmers are an expensive resource if they spend too much time waiting for their jobs to run. /snip In other terms: If I have a STC consuming 25,000 cpus / secs per month, and I spend two days optimizing it to reduce to 15,000. Am I gaining money or loosing time ? /esnip Getting back to your first question, I'd say maybe. If you need that 15,000 CPU seconds for other work, or if trimming them keeps your peak down for the month, you may save more than would be obvious. Your customer may be delighted that you care about keeping their bill down and may bring you a bunch more billiable work. It depends. HTH, Linda ----- Original Message ----- From: "Salva Carrasco" <scarra...@unicaja.es> To: IBM-MAIN@bama.ua.edu Sent: Tuesday, June 12, 2012 2:13:24 PM Subject: Re: How many cost a cpu second? Apart from the syntax error "How many" vs. "How much", My question is not having much success. In other terms: If I have a STC consuming 25,000 cpus / secs per month, and I spend two days optimizing it to reduce to 15,000. Am I gaining money or loosing time ? ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: INFO IBM-MAIN
