As long as you run EXACTLY the same thing on each linux server with the same users, same transactions and same configuration, then you can be successful in answering how many will fit on an IFL. :-)
Our charge-back model is that there is a basic monthly charge for the server and support, tiered by what level of support and high-availability you want. I believe there are three or four basic server "models" you can select from; from non-HA with "when we get to it" support to full HA w/DR and 24/7 support. All of those include a basic amount of CPU time, memory and storage. Anything added/used over and above those basics are additional charges. For instance, we include 15 MIPS for zLinux servers in the base price. If they want to be silly and leave jvms running that are looping or poorly written, and the server consumes 50 MIPS every day, they get charged more for the 35 extra MIPS. The ones that are "funny" are the loopers; they leave apps looping for days on end and then wonder why their bill is so high. The charge-back does help them understand that they need to pay attention to their applications more. We also charge per 256M of virtual memory defined over the base. Storage is a flat rate per MB or GB and is no different than any other platform. -- Jim Vincent On Wed, Nov 18, 2009 at 1:02 PM, Dave Keeton <[email protected]>wrote: > What a timely question! Management is after me to quantify how many guests > we can run on our 1 IFL for billing purposes. They want to carve it up into > units so they know how many total we have, and therefore can charge per > server. I don't see it going that way, as several of you have pointed out: > it depends on the workload. > > Anyone running in a data center and charging for their Linux servers? How > are you approaching it? Do you have a flat fee per instance, or do you have > a charge-back model? I'm curious and completely new at this aspect of z/VM > management. > > Thanks, > Dave >
