立即与支持人员聊天
与支持团队交流

Foglight Chargeback 5.7.7 - User Guide

Pros and Cons of Using the Tiered Flat Rate Model

The Tiered Flat Rate model assesses costs as a monthly flat rate. As the costs are calculated in this manner, changes to tier definitions take place immediately and can be applied in a retroactive manner. For example, running a Tiered Flat Rate report with the tiers configured in different ways will generate different costs against the same historical time period.

When using the Tiered Flat Rate model, a department is charged a flat rate for their tier of VMs (for example, Tier 1 is one vCPU, $100/month) and know what their costs will be each month. The drawback of using the Tiered Flat Rate model is getting departments to agree on tier rates. For example, if the VM is turned off, the department is still charged.

Measured Resource Utilization Model

The Measured Resource Utilization Model (MRU) can be thought of as the charges you would get from the power company; you are charged for what you use. With the MRU model, each VM is only charged for what it uses of its server's resources. MRU calculations are performed by determining each virtual machine's actual utilization and applying the appropriate cost-unit calculation for each resource based on the resource weighting configuration. The MRU model provides methods for calculating costs based on the following types of usage:

Understanding MRU Chargeback Calculations

Foglight Chargeback runs MRU hourly background calculations. When the calculation runs, it inventories where all of the VMs are on each server and assumes, (for calculation simplicity, ignoring VMotion events) that each VM has been resident on that server host for the entire hour.

The hourly calculation determines both full cost recovery and partial cost recovery charges automatically. You cannot turn it off or adjust it in any way. The calculation is based on the weights (percentages) for the server CPU weight, memory weight, network weight, and Disk IO weight used by the VMs resident on that server.

The MRU calculations do not begin to generate until after the servers are assigned to the templates. At the next hour, the background calculations are performed and continue from that point. The calculations are based on detailed current usage assessments; the MRU calculations are not retroactive. The historical data may not provide the precision needed to generate accurate hourly costs of the past usage.

Similarly, if any change is made to an MRU host template, these changes take effect beginning with the next hourly cost calculations. These changes are also not retroactive because the costs have already been calculated based on the previous configuration.

MRU templates should be configured to calculate the charges needed to recover the cost of an individual server even when multiple servers in the infrastructure have identical hardware. Each server that is assigned to a template individually recovers the costs outlined in the template definition.

An organization can simultaneously and indefinitely run reports using the Tiered Flat Rate + Add Ons (see Add Ons Model) or MRU + Add Ons models to compare the chargeback costs calculated using each model. Since the organization ultimately needs end user department buy in, the chargeback model(s) are straight-forward and fair. At some point an organization may decide to use TFR + Add Ons or MRU + Add Ons.

Full cost recovery means that a department is charged for their VM usage of the server in order to recoup all the costs of operating that server. Full cost recovery involves an organization where the information technology (IT) departments cannot be a cost center. IT cannot be liable for servers and infrastructure costs. All IT costs must be fully recovered by charging customers or business units. With this chargeback method, the number of virtual machines (VMs) running on a host has an effect on the calculated cost for each VM. As a simple example, in the first month a $5000 host has five VMs all performing identically across all resources. The overall cost is approximately $1000 per VM to fully recover the cost of the infrastructure. The next month, IT has added five VMs, bringing the total to ten. Now each virtual machine costs $500 to recover the cost, which is half the cost of the previous month.

In this example, the cost for each VM originally came out to $1000 based upon even usage across the VMs and that the sum of the costs of the VMs needs to reach the $5000 monthly cost of the server. When the additional VMs were added for the second month, the costs come out to $500 per VM at the end of that month, keeping in mind that the sum of the VM costs still add up to the $5000 monthly cost for the server. The costs for the first month remain at $1000 per VM. These costs were calculated based on the VM usage and server template assignment at that time.

This method eventually balances itself out over time as the virtual infrastructure reaches an acceptable capacity. This helps virtualization administrators provide a balanced virtual infrastructure of cost and performance, providing IT Management with the capability to drive proper hardware utilization and prevent over-allocation of virtual hardware.

For partial cost recovery, a department is charged for their use against the capacity of the server. Partial cost recovery is used in instances where IT absorbs costs for portions of the infrastructure that are going unused. The costs calculated for the VMs adjust based on the overall utilization of the server against its capacity. This allows the IT department recover costs for the load that each department puts on the infrastructure and only use the IT department budget for the cost of the unused capacity.

Example of Measured Resource Utilization Cost Calculation

The following prerequisites are used:

The following example shows the template used to calculate the Measured Resource Utilization model. The user enters the total server cost, the lifecycle of the server, the resource weighting based on a percentage model that considers the core four resources:

You can specify any percentage value to the core resources. If the values sum is more than 100 percent, IT recovers more than the cost of the server. If the values sum to less than 100 percent, IT will not recover the total cost, even in Full Cost Recovery mode.

MRU templates are assigned to VMware (ESX) or Hyper-V servers. MRU Full and Partial Cost are calculated for the Virtual Machines hourly, based on the template assigned to the server they reside on.

The following provides the example data entered:

相关文档

The document was helpful.

选择评级

I easily found the information I needed.

选择评级