Understanding VRB calculations

VRealize Business is VMware’s Cloud cost analysis and showback tool. While VRB is around for long there is a lot of confusion around how the calculations are done in VRB.

With this article I hope to address the basic per VM pre month cost calculations.

For calculation purpose VRB uses few cost drivers, some of these cost drivers cost is monthly while some is aggregate cost. Let us first understand the cost drivers and other inputs used for the calculations

Cost drivers :- VRB uses 8 basic cost drivers as listed below

  1. Cost of server hardware
  2. OS licenses
  3. Labor for managing physical servers, OS and virtualization
  4. Facilities
  5. Storage
  6. Network
  7. Maintenance fees of HW and OS
  8. Additional costs such as backup, disaster recovery, security, etc.

 

Step 2 :- we calculate the depreciation cost of hardware, the formula for the same is as below

Yearly Double declining depreciation: (original cost – accumulated depreciation) * Depreciation Rate

Yearly Straight line depreciation: (original cost – accumulated depreciation) / Number of remaining years to depreciate

Yearly Depreciation: Max (Yearly Double declining depreciation, Yearly Straight line depreciation)

Note:- refer the xls sheet for example.

In a nutshell, we select the max depreciation value between Double declining and Straight line as depreciation value for the hardware

Step3 :-  Now we calculate the effective CPU and Memory using the Expected utilization.

Expected utilization is the what is the usage % per host we want to achieve, this can be dictated by RFP or SLA. For example we can have a host with 1 TB of ram, but we need to leave 30% as buffer. So the expected usage becomes 70%.

Use the following formula to calculate the effective CPU and Memory,

For CPU :-

𝐸𝑓𝑓𝑒𝑐𝑡𝑖𝑣𝑒 𝐺𝐻𝑧 = 𝐸𝑥𝑝𝑒𝑐𝑡𝑒𝑑 𝑈𝑡𝑖𝑙𝑖𝑧𝑎𝑡𝑖𝑜𝑛 × sum ([𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝐶𝑃𝑈𝑠 × 𝐶𝑜𝑟𝑒𝑠 𝑝𝑒𝑟 𝐶𝑃𝑈 × 𝐺𝐻𝑧 𝑝𝑒𝑟 𝐶𝑜𝑟𝑒]) for all hosts

For RAM :-

𝐸𝑓𝑓𝑒𝑐𝑡𝑖𝑣𝑒 𝐺𝐵 = 𝐸𝑥𝑝𝑒𝑐𝑡𝑒𝑑 𝑈𝑡𝑖𝑙𝑖𝑧𝑎𝑡𝑖𝑜𝑛 × sum(  [𝑅𝐴𝑀 𝐺𝐵]) for all hosts

Step 4:- Next we calculate the fully loaded cost of the host and cluster,

The fully loaded cost of cluster is sum of

Server Cost (considering depreciation) + Network + Facilities + Maintenance

To find the cost of fully loaded cluster, we do a summation of fully loaded cost of all the hosts. This value will decrease every year due to depreciation cost.

Step 5 :- Now we need to find the relative CPU and Memory cost of the cluster using the total cost of the cluster.

This value does not contain the storage cost in traditional infra. But in HCI we need to consider the storage cost aswell.

To find relative cost of CPU and Memory, we will consider how much percent of cost was CPU cost and how much is memory cost.

Step 6 :-  Now using relative cost of the CPU Memory and Storage we will calculate,

Cost per GHz :- relative CPU cost / effective CPU

Similarly, Cost per GB of ram :- relative per GB cost / effective GB

Cost per GB of storage :- relative cost per GB storage / effective GB storage (only in case of HCI)

Step 7 :- now using these values we can just add up the cost per VM.

Eg:-        Cost per GHz = 6

               Cost per GB ram = 4

               Cost per GB storage = 1

 

For a  “4 vcpu , 8 GB ram, and 500 GB storage” cost per VM = (4*6)+(4*8)+(1*500) = 556

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s