Chargeback on Orion Based VMAN Install

princeal  wrote:

This is a feature that I would very much like to see. Is there a way we could still run the legacy version of VMAN to get this feature back?

You can certainly keep running the appliance and have it integrated with VMAN on Orion to utilize this feature. That's still supported. VMAN 8.4 continues to integrate with the 8.1 version of the appliance.

This is a feature that I would very much like to see. Is there a way we could still run the legacy version of VMAN to get this feature back?

horizenred  wrote:

I spoke extensively with a peer today to see how we would best approach this and how to best reply to you.

Since we would not be actually implementing chargebacks, but are looking at this at a showback (cost analytics) level, and since we are not in the business of IT as a service, we decided that we would want to price things simply as if we would earn our money back at the end of the equipment life cycle. In regards to your question above, I believe that we would charge compute of a VM based on a per cpu or per Ghz allocated to the VM. Since there is no profit to be made, there is no concern on cycles consumed. Memory would also be charged as allocated.

Storage would be charged by data consumed. The only consideration here being if the VM storage was thick or thin provisioned. If thin, then only charged for the total space used by the end of that billing cycle. If thick, then always billed for the full amount. The thin provisioning would allow for you to see the cost growth month to month which could be useful.

These three numbers would at least get you the hardware costs per server, which doesn't take into account networking, electrical consumed, cooling, additional HW licensing (like VMWare licensing), administrative costs (labor), monitoring software, etc.

This is great, thank you for sharing this insight.

I spoke extensively with a peer today to see how we would best approach this and how to best reply to you.

Since we would not be actually implementing chargebacks, but are looking at this at a showback (cost analytics) level, and since we are not in the business of IT as a service, we decided that we would want to price things simply as if we would earn our money back at the end of the equipment life cycle. In regards to your question above, I believe that we would charge compute of a VM based on a per cpu or per Ghz allocated to the VM. Since there is no profit to be made, there is no concern on cycles consumed. Memory would also be charged as allocated.

Storage would be charged by data consumed. The only consideration here being if the VM storage was thick or thin provisioned. If thin, then only charged for the total space used by the end of that billing cycle. If thick, then always billed for the full amount. The thin provisioning would allow for you to see the cost growth month to month which could be useful.

These three numbers would at least get you the hardware costs per server, which doesn't take into account networking, electrical consumed, cooling, additional HW licensing (like VMWare licensing), administrative costs (labor), monitoring software, etc.

I've thought on this a bit as well, and am interested in how others calculate the compute cost with regard to different chipsets: would you determine the cost based on CPU model, generation, etc; the cost based on SPECInt numbers, or just a rough and rounded cost by MHz?