Pricing for IBM Power Virtual Server (On-premises)

Monthly usage

IBM Power Virtual Server (On-premises) instances are charged at a monthly rate that is prorated per hour. If you add resources to a virtual machine in the middle of the month, the bill for the virtual machine reflects the resource change at a per-hour prorated price.

All prices that are mentioned in the topic, Pricing for IBM Power Virtual Server Private Cloud, are illustrative and do not represent the actual amounts that are used for billing. To generate an estimated price, use the Power Virtual Server Estimate pricing tool. For more information, see Generating an estimate.

In the following example, the customer provisions an IBM Power Virtual Server (On-premises) instance. The instance has 1 core with 8 GB of memory, a 150 GB disk, and is running Red Hat Enterprise Linux (RHEL) operating system (OS). Let us assume the following hypothetical monthly prices:

• Cores: $143.23/month x 1 = $143.23
• Memory: $9/month each GB x 8 = $ 72.00
• Storage: $0.216/month each GB x 150 GB = $ 32.40
• OS license: $43.80/month each core = $ 43.80

Total cost per month = $291.43

Factoring the example quantities and prices, the base price for the virtual machine is $291.43 per month or an average of $0.399 per hour for all the resources consumed. Assume that in the course of the month, the customer allocates more memory to the virtual machine. At the end of the month, the new price for the virtual machine will be higher to account for the extra memory. The new price amounts to $0.498/hour for all the resources consumed. The monthly bill is prorated by an hour for the resources that are deployed as shown in Table 1.

In Table 1, the virtual machine resources are increased (after the resources reach 300 hours in the month) from 8 GB to 16 GB of memory. The price of the virtual machine is prorated by an hour for the final monthly price of $303.

For the example in Table 1, we might see four distinct part numbers for the following resources in the invoice:

• The virtual cores (one for each core)
• The memory
• The volume storage (disk)
• The OS license (RHEL)

For detailed usage and billing information, refer to the part numbers in your invoice. The part numbers in the invoice represent the charge unit for each resource charged. To view the part numbers, refer to Table 2.

For more information about unit prices for each metric ID, see Where can I find the unit prices for the billing metrics?.

Base instances

The billing of a base instance depends on the options you select when you create a virtual machine. The machine type, number of cores, and amount of memory affect the base instance billing. When you create your virtual machine, the associated monthly rate is displayed on the billing of the base instances.

The pricing for memory is calculated based on a ratio of 64 GB per core. For example, if you use more than 16 GB for 0.25 cores, you must pay a premium high-use RAM price for the excess memory. However, if you use up to 128 GB for 2 cores, you do not have to pay any premium memory price.

Operating systems

Linux and AIX operating systems are supported and only RHEL stock images are available. For more information about supported versions and distributions, see Full Linux subscription for Power Virtual Server (On-premises). When you select a RHEL Linux stock image, the pricing includes the full Linux subscription charges. The full Linux subscription allows you to subscribe to updated or upgraded packages from a Red Hat Satellite server by configuring the provisioned virtual machine. The usage meter for RHEL full Linux subscription starts when you create the virtual machine. If you Bring Your Own License (BYOL), it is not metered or billed.

If you bring your own custom image, you are charged for the image size and the storage tier that you use for the custom image. After you deploy a stock image (and only after deployment), you are charged for the space that is used to store the stock image. The storage unit price (per GB) for the stored boot images is the same as the selected storage tier (Tier 0, Tier 1, or Tier 3) where your boot disks are deployed. To estimate the storage rates, use the Power Virtual Server Estimate pricing tool. To reduce costs, select the virtual machine that is not needed and delete it.

Processor types

You are charged different rates depending on the type of system that you choose for your virtual machine. Enterprise systems, such as E1080 or E1050, cost more because they provide more processing capacity per core. For example, the E1080 costs more than E1050.

You can choose one of the following core types for your workload:

• Dedicated virtual processor cores
• Shared capped virtual processor cores
• Shared uncapped virtual processor cores

Different scenarios provide different benefits when you use each type of virtual processor core.

All prices that are mentioned in the topic, Pricing for IBM Power Virtual Server Private Cloud, are illustrative and do not represent the actual amounts that are used for billing. To generate an estimated price, use the Power Virtual Server Estimate pricing tool. For more information, see Generating an estimate.

In the following example, assume that the cost of different types of virtual processor cores is the same within each system type they belong to.

Tables 3 to 5 show how different processor types affect the cost per system:

Memory types

Charges for IBM Power Virtual Server (On-premises) are determined by the memory usage (in gigabytes), and the type of the system chosen for the virtual machine. Enterprise systems, such as E1080 or E1050, use scale-up memory with higher density, whereas S1022 systems use scale-up memory with lesser density. Currently, both scale-up and scale-out memory might cost you the same price. However, this pricing is subject to change in the future, depending on the cost of the memory chips. Table 6 shows how memory types affect the cost per system:

Pricing for dedicated hosts

Dedicated hosts are priced based upon the host type – either an IBM Power S922 or IBM Power S1022. Each server type is metered by the hour and the price includes the entire capacity of the host.

Consider the following points for dedicated host pricing:

• You are not charged separately for shared processor pools you deploy to the dedicated host.
• Software charges for the supported operating systems are metered and charged by the core.

To learn more about the dedicated host, see: dedicated host.

Storage types

IBM Power Virtual Server (On-premises) charges are based on the following storage types:

• Data volumes: These volumes are basic forms that you can create. The billing depends on the volume size when metered. Table 7 shows an example of how you are billed based on your volume creation:

  Table 7. Calculation of data volume

  Volume size you create	Volume size you are billed
  10 GB	10 GB
  10+5 GB	15 GB

• Image backing volumes: These volumes are part of a boot image in your cloud instance listed in the boot image catalog. Billing for image backing volumes is determined as follows:

  • Single volume billing - When the image has a single backing volume, you are billed based on the size (GB) of the single volume.
  • Multiple volume billing - When the image has multiple backing volumes, billing is based on tallying up the sizes of all the image backing volumes.

Table 8 shows an example of how you are billed based on your boot volume:

• Deployed virtual machine volumes: When you deploy a virtual machine with an image, it gets a copy of all the volumes in the image. Extra data volumes added to the deployed virtual machine are billed separately as Data Volumes. Table 9 shows an example of how you are billed based on the virtual machines that you deploy:

  Table 9. Calculation of deployed virtual machines volume

  Image backing volume	Billed volume size
  20 GB	20 GB
  20 GB + 30 GB	50 GB

• Deployed virtual machine snapshots: The snapshots of the volumes are taken after the virtual machine is deployed. The size of the volume snapshot is related to the number of updates that are made to the virtual machine. When you take a snapshot for the first time, the size of the snapshot is a fraction of the size of the volume(s) of the virtual machine. The size of subsequent snapshots might increase based on the changes that are made to the original volume.

For an example, consider a virtual machine with a volume of 100 GB. The size of the first snapshot is 100 GB. The size of the second snapshot might be 1 GB.

Snapshot sizes are not predictable as they are related to the updates made to the volume between two snapshots.

Tables 10 and 11 show how different storage types affect the cost per system:

End of billing

The monthly billing cycle ends when you delete the virtual machine. If you scale your infrastructure up and down in response to workload requirements, your billing follows the timing of the virtual machine provision change. If you stop the virtual machine, the billing process is not stopped. Delete the virtual machine to stop the billing cycle.

When your virtual machine is active but idle, you can use Dynamic Logical Partitioning (DLPAR) to resize it to a minimal state. You can drastically decrease the price per hour by reducing the core count and memory of the virtual machine.

Frequently asked questions

Review the following frequently asked questions about pricing: