In our set up we have mixed both IBMi and OpenVMS to the sub-tower Midrange but I wonder if we are on the right track. Have you done the same? If not, where did you put OpenVMS?

In the Benchmarking, we didn't get any valuable numbers, as we could not convert the CPUCores to an accurate number of MIPS, but for IBMi we could convert the CPW to MIPS.

All input is appreciated.

No idea re OpenVMS, but for iSeries, we have that under Compute-Mainframe (we're still on v1 of the taxonomy).  We have those servers in our CMDB and really, for that particular allocation line, apps receive the cost and it's weighted based on number of servers the apps appear on.  If they show their platform as iSeries in the app inventory, but nothing shows in CMDB, then they get what remains of the iSeries dollars weighted based on 'average transaction volume' (also in the app inventory). 

I know you didn't ask about zSeries (also Compute-Mainframe), but our Mainframe team went through the laborious process a while back of mapping mainframe programs to the apps in the app inventory.  They send us a file each month that shows percentage of CPU usage per app, so we use that to weight the dollars.  The total percentage comes through around 97-98 percent each month, so those apps that are within that file get their dollars based on the percentage provided.  The 2-3 percent that remains is spread to any apps that are not on that list, but are in the app inventory showing a platform of zSeries (also spread based on 'average transaction volume').  

Went off on a tangent, there, but semi same vein - sorry for not having any feedback/info on OpenVMS, though.

Thanks @Jenny Franklin. We're on v2 of the taxonomy, and there IBMi is mapped to Midrange. We're not a huge player in the IBMi ballpark, so we have a fixed file we review manually every quarter. 

Interesting to hear how you handle zSeries. We do it a similar way, but we already had the information on how many CPU-seconds each and every application was consuming every month (from our internal invoicing system) and could with that information split the cost for Mainframe. Had quite an easy ride there :-)

We also tried to get some input from Apptio regarding other parameters to use (SMF-records) in the non-existing insight reports, but I have understood that Mainframe is not on the top of the agenda. Pitty, as it would be valuable for many of us. :-(