Public Cloud Pricing

Sam Kirchoff

Running workloads in the “Cloud” means different things to different people. This entry is focused on the Public Cloud and, more specifically, the big three: AWS, Azure, and Google Cloud.

It’s a common scenario that users will migrate or place a workload in the Public Cloud only to find out the cost is a little hard to stomach. This is often due to lack of proper planning knowledge upfront. So naturally, Live Optics set out to make it easier for IT professionals to understand this and keep it “point and click” easy.

Before I continue, I want to thank and old friend and fellow technologist, Rick Gouin, up at Winslow Technology Group in Boston for giving us early feedback on our work.

Companies such as Amazon, Google, and Microsoft have taken the approach of making pricing publicly accessible for their Public Cloud products through the use of online calculators or through their APIs. While very convenient, the average user might not want to spend the hours scripting the APIs or manually entering in values to a pricing calculator.

These public cloud providers price VMs by way of matching them against a matrix of configurations. To accommodate the variances in CPUs, Memory, and Capacity these matrixes are very large, often in the thousands of configuration options.

Accurate planning means accurately inputting data into these tools. Often IT professionals don’t have inventory of their VMs or don’t want to go through the process of entering all this data into an online calculator. And while one might be stubborn enough to hand-enter a few hundred machines into one of these vendor’s tools, the chances of them repeating that process across multiple vendors to “shop” the right technology partner would be low.

Live Optics helps you solve both problems. First, Live Optics will automate the inventorying of your VMs with just a few clicks. Second, Live Optics has now introduced automatic cloud pricing and comparison for every VM inventoried in your project.  

What’s the Goal?

Let’s talk about what this IS and what this IS NOT.

In its current state, this feature is not designed to be a precision calculator of Public Cloud cost. Live Optics set out to provide directionally correct -- yet still accurate -- pricing based commonly accepted inputs between these top three Public Cloud Providers.

There are dozens, if not many more ways to customize spend in the public cloud. Some of these will drive the cost up and some will drive the cost down.

By settling on these common inputs, we can easily find a way to peer-compare the three providers. Pricing is rather consistent but does have fluctuations based on VM configurations and there is never one clear winner.

Pricing, even with all things being submitted in a consistent manner, gives us a neutral piece of information to help us drive insights on more than just costs alone.

For example, if we wanted to calculate a 95% cost outlier, we’d find this very straightforward once we have a baseline price established.

As a second example, what if we wanted to know all the VMs that were above the average cost of a VM on a platform?

Or lastly, what if we wanted to know which VM configurations currently do not fit any configuration options provided on these platforms!

You will see as we expand this offering there is far more that can be derived from price than cost alone.

What is being submitted?

Currently, Live Optics has established a baseline of assumptions that go along with pricing a VM. The good news is that if someone replicated the exact same inputs directly to an online calculator, they would get the exact same answers. Let’s review what those are in detail.

Shared Tenancy:
While it is possible to get a dedicated hosting agreement, in its current implementation Cloud Pricing is assuming that one will be estimating for Shared Occupancy pricing results.

Location:
The current location of hosting is fixed at “US East” as a baseline. Those that are in different geographies will need to account for local pricing.

OS:
The type or variant of OS that you use in the Public Cloud will drive price. Also, its inclusion will drive price.

The Cloud Pricing feature will assume that OS rental is needed and that a Bring Your Own License (BYOL) model is not being pursued.

The closest match to the OS from the VM configured in your inventory will be used. Some Linux variants might not have matches and therefore will be assigned a generic Linux license.

Virtual CPU:
The number of currently assigned vCPUs to the VM in your inventory will be submitted for CPU matching. It’s common that privately hosted virtual environments might have vCPU assignments that aren’t attainable in the public cloud. When no match can be found, the largest closest match will be used.

Memory and Capacity:
Memory and Capacity values are also used from the VM definition inventoried. There is a +/- 10% margin of error applied here to help find the most compatible match. It should be known that most Public Cloud providers will round up to the next highest match. This is important as small overlaps in the pricing matrix can lead to expensive monthly costs that are completely avoidable.

SSD Storage:
It is assumed that SSD storage is the desired capacity assignment.

Other Fixed Variables:
Live Optics has also fixed the currency to the US Dollar. There is an assumption that the VM will be running 24/7.

Note: please see the release notes for more logic on roundups for CPU, Memory, and other smart approximations at https://support.liveoptics.com/hc/en-us/articles/360051474753

Discount Costs for Time Commitment

In the migration from physical to cloud, most will go from physical to virtual first -- typically, on premise. It’s not common and sometimes not even possible to go from physical directly to the cloud. For this reason, Live Optics only prices out the VMs in a project.

Similarly, it would be very uncommon for someone to agree to an annual or even a 3-year commitment contract without some trial period.

This leaves most first time Public Cloud users experiencing the “On Demand” pricing, which is essentially month-to-month billing and is obviously the costliest way to consume these services.

For planning purposes Live Optics includes the On Demand pricing, but also includes the pricing for 1- and 3-year commitments.

Moving or Repatriating VMs

Obviously, this Cloud Pricing feature is there to help you understand the price of moving to the cloud. However, what if you have workloads that are already there?

With the Live Optics server/virtualization collection model, you aren’t limited to collecting VMs and workloads that are just on premise or even on a specific vendor.

The collection is vendor- and platform-agnostic and will work the same in the Public Cloud as it would locally in your data center.

If you are looking for a great way to detail VM workloads that you seek to move between Public Clouds or even back on site, then Live Optics has a turnkey solution for this exercise already.

What you get back

The onscreen display will do a quick analysis for you by adding up all the 1 year discount pricing for each VM. This is performed for each Public Cloud Provider so the pricing can be compared in near real-time.

It will also calculate the average VM specifications and show you the monthly cost at the 1 year discount that was returned from each provider.  

ScreenPCP_small.jpg

Was this article helpful?
0 out of 0 found this helpful
Please sign in to leave a comment.