Re: Chargeback for zIIP & zAAP Usage
Al Sherkow writes: >If you lower the general CPU rate you are passing on some of the platform's >lower TCO with zXXP to *all* the users/customers. >If you carry this far enough you lower the general CPU rate, and then you >could use your option "A. The same as for regular CPU."! If you believe chargebacks should accurately reflect true costs, I don't think I would agree with a single "blended" rate (for the same reasons I don't think there should be a single rate blending fixed and variable costs). End users do have the ability to influence zAAP and zIIP eligibility. If they fail to upgrade from backlevel Java, then they won't get zAAP benefits and they shouldn't pay the same rate as another user who did go to the trouble of upgrading, to pick one example. Or if they copy (stale) data from DB2 to some other data store via FTP rather than just accessing it directly via JDBC or ODBC, there should be a real reflection of true costs and the very different zIIP results between those two cases. But I do see your point about zAAP and zIIP eligibility versus actually running on the specialty engines. If zAAP or zIIP eligible work gets dispatched to the CPs, what rate should that work be? Said another way, do the operations people have any responsibility for trying to maximize the amount of eligible work that actually gets dispatched to the specialties? How do you preserve the incentives for operations to be as efficient as possible? I think the typical way business users keep their IT staff "honest" is by checking the price of outsourcing periodically -- getting competitive bids and comparing them to their internal organization, basically. This too raises a number of questions and potential problems, but it is one method. I generally find that bad chargebacks are worse than no chargebacks, so if you have to pick one pick the latter. "Worse" means they do more damage to the company's own financial and innovative interests. And there are a lot of bad chargeback regimes out there, unfortunately. For example, a lot of companies still put most or all of the data center facilities costs into mainframe chargebacks. That's just nuts and totally indefensible in 2008: the mainframe is the smallest and least demanding piece of equipment in many data centers. - - - - - 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
Re: Chargeback for zIIP & zAAP Usage
Ted -- I also would lean towards "C. Some sort of discount". I also think the rate for general CPU could be lowered due to zXXP engines (I borrow Kathy Walsh's shorthand). Your *users* do not necessarily have control over whether there work goes to zXXP or not. Some of this depends on how many zXXP you have installed and available in LPARs relative to the amount of work that could use the zXXP. If you lower the general CPU rate you are passing on some of the platform's lower TCO with zXXP to *all* the users/customers. If you carry this far enough you lower the general CPU rate, and then you could use your option "A. The same as for regular CPU."! Al Sherkow, I/S Management Strategies, Ltd. Consulting Expertise on Capacity Planning, Performance Tuning, WLC, LPARs, IRD and LCS Software Seminars on IBM SW Pricing, LPARs, and IRD Voice: +1 414 332-3062 Web: www.sherkow.com -- 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
Re: Chargeback for zIIP & zAAP Usage
Answer: C. Some sort of discount. It should also be a fixed-plus-variable regime. The marginal cost of one additional "unit" of processing on a zIIP or zAAP is trivial, so if you charge the average cost you convey all sorts of perverse incentives to users. I also tend to think that different service classes ought to have different charges. The costs should be declining over time as well. zIIPs and zAAPs get faster with a model upgrade, for example. - - - - - 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
Chargeback for zIIP & zAAP Usage
Has anybody looked into this? How are you billing, if using chargeback? Since these specialty engines are designed to reduce software costs, should you charge: A. The same as for regular CPU. B. The cost based on the engine's 'true' power. C. Some sort of discount. I tend towards 'C', but am curious as to what others do. PS: I'm not a big fan of Chargeback; managers are. - Too busy driving to stop for gas! -- 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