>>> On 9/7/2012 at 06:56 PM, Mike Schwab <[email protected]> wrote: > 10 or 20 Linux servers consolidated onto 1 x86-64 blade server. > 300 Linux servers consolidated onto 1 zIFL. > > Now that looks reasonable. A full speed z processor is still 15 to 30 > times faster than Virtual x86-64.
Um, no, that's not what those number mean. A full speed zEC12 processor at 5.5 GHz, is only about 53% faster than a 3.6GHz Intel processor. It's possible to consolidate a very large number of mostly idle systems onto a single System z processor. There are other systems running workloads that require 1 or 2 or even more System z processors for a single system image. This is the point that most people just don't seem to want to understand. What platform is best for a given workload really does depend on the characteristics of that workload, and the business requirements that surround it. There is no one platform that is best for all workloads. Period. Anyone that tries to tell you one platform is the best for all possible workloads is either lying or incompetent. Anyone that tries to tell you that the purchase price of the box divided by the total number of instructions executed per second is the only metric to consider is either lying or incompetent. Doing real TCO comparisons between disparate architectures is hard and a lot of work. Because it's hard, most people don't bother trying to do a very good job at it and make overly simplistic models that provide the answers they want to hear in the first place. Mark Post ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
