Editors Note: Updated October 2, 2024
Three years ago, Snow Software launched a SaaS platform, Snow Atlas, to help their customers manage their hybrid-IT technology landscape. Because this is a relatively new offering, finance and the CIO were staying on top of tracking the overall cost of the cloud bill and understanding unit economics as these influence the product go-to-market strategy.
Snow and Flexera are active members of the FinOps Foundation, driving forward standards and best practices for the convergence of ITAM and FinOps, and many of our staff are FinOps Certified.
Barna Tarczali, Snow Atlas Engineering Manager, is one of the key representatives of the FinOps team because of his expertise in developing solutions to track license consumption in cloud environments.
The Snow Atlas platform leverages Microsoft Azure for infrastructure and hosts this application in four regions including the UK, America, Australia and the Netherlands. When looking at the Microsoft Azure portal, the team saw Snow was spending $2M annually on VMs in Azure. With Barna’s expert knowledge of FinOps practices, he knew this spend not only included the cost of the cloud infrastructure, but also included software costs.
Barna proceeded to investigate the breakdown of IaaS and software license costs. Finance and Engineering teams are often entirely unaware of the breakdown, not realizing they could be spending a significant amount on pay-as-you-go (PAYG) license costs that could be replaced with a more cost effective license including BYOL licenses.
Using the Azure Cost Management, Barna found that 25%, or half a million of the $2M Azure bill was for software license costs, primarily from PAYG licenses for Windows Servers.
Microsoft Azure offers payment options such as the Pay As You Go (PAYG) model, where charges are based on usage. While this provides flexibility to scale licenses as needed, it comes at a higher cost compared to other licensing models like Bring Your Own License (BYOL). Barna was aware of Microsoft’s BYOL program called Azure Hybrid Benefits Use (AHB) and realized that they were not utilizing this cost-saving option for their VMs.
To determine which licenses to move to the BYOL option, Barna took the following steps:
- First, he identified the areas of highest spend to get the most savings. In combing through the data, they found the Kubernetes Windows Servers made up approximately 66% of the total cloud license costs.
- Next, they looked to understand the types of workloads on these Kubernetes clusters. License optimization using BYOL requires a minimum 1-year lock-in, so if these workloads were anticipated to be removed soon, optimization would yield minimal benefit. Barna confirmed that out of the 65 instances, 80% would remain in operation for the near future.
- Determining which nodes needed optimization and where BYOL could be applied is a crucial decision requiring collaboration between the FinOps and Engineering teams. Barna checked with the software asset management team and found there were no unallocated Windows Server licenses available. They decided to procure additional licenses to reduce the PAYG costs and started first by transitioning 32 Windows Server licenses from PAYG to BYOL, the equivalent of 512 core licenses.
- Snow approached their Microsoft license provider to obtain BYOL license options for converting these 512 Windows Server Core Licenses. One option was to move to a subscription model with a 1-year commitment and the other option was purchasing perpetual licenses with a 3-year term with included Software Assurance. Based on risk profile and financial preference, Snow team decided to pick the subscription option which resulted in a cost savings of >90% for these 32 Windows Servers.
| License Type | Annual License Cost for 1 VM | Annual License Cost for 32 VMs |
| PAYG | $6,302.92 | $201,693.44 |
| BYOL | $597.65 | $19,124.80 |
Annual Savings = $182,568, or 90.5%
Return on Investment = 1.14 months
More savings, more responsibilities with BYOL
Moving licenses from PAYG significantly reduces spend but adds the responsibility of ensuring license compliance, especially when workload requirements increase and adding additional servers is needed to meet customer demand. Monitoring license requirements using the native Azure portal is a manual task, necessitating the FinOps or Engineering team to filter by VM type, and calculate instances to determine required licenses.
This process becomes intricate with different VM types. For instance, if you have seven “Standard F16s v2” instances, you would multiply 16 (number of cores per instance) by 7 (number of instances), resulting in a total of 112 core licenses required. Similarly, if you have 8 instances of “Standard D8s_v3,” then you need to multiply 8 (number of cores per instance) by 8 (number of instances) for a total of 64 core licenses. These calculations cannot be easily seen on your Azure portal and need to be done in Excel to manage cloud licenses.
Click to view full screen.
Additionally, most FinOps and Engineering teams are not license experts nor do they typically have access to enterprise agreements. This is typically the responsibility of software asset management (SAM) experts in the organization. However, the software asset management team normally has no visibility to the cloud bills, most times spanning across vendors. In large organizations, there can be several logins for various cloud environments.
In short, on both sides of the FinOps and SAM fence, there is a visibility gap to understand license costs, waste and compliance.
According to the FinOps Foundation’s State of FinOps Survey, organizational alignment is the biggest challenge to achieving goals. And a high priority goal for both ITAM and FinOps practitioners is to reduce waste.
This is exactly why Flexera is focused on solving the hybrid ITAM and FinOps visibility gap.
Time savings with automated BYOL analysis and license compliance monitoring
While the Azure Cost Management tool helps you understand where you are spending on PAYG software, as outlined above, there are a lot of manual steps to understand:
- Are there available licenses from our existing pool that can be used for BYOL eligible workloads?
- What is the calculation of potential savings for switching from PAYG to BYOL?
- License compliance between on-premises and cloud workloads.
Barna’s journey of saving 25% of the Azure bill was for just one service. Imagine an organization that develops hundreds of services, leveraging multiple cloud environments. Manually researching available licenses, potential savings and license compliance could take weeks of work. When work is manual, and no one’s full time job, the task falls to the bottom of the to-do list.
To reduce this manual effort at scale, and help organizations save 5 or 10% of their cloud bill, Flexera has developed a solution to shine a light on the blind spot between FinOps and ITAM teams – software license costs in the cloud.
The Cloud License Management feature, available in Snow Atlas, provides instant visibility to understand:
- What Windows Server and SQL Server workloads are BYOL eligible?
- Are there available licenses you already own?
- What is the savings impact of moving from PAYG to BYOL?
- What is your effective license position across your on-premises and Azure environment?
Below is an example illustrating how Snow Atlas highlights cloud instances that are BYO- eligible and the estimated savings for moving to BYOL from PAYG. We also track how much you’ve already saved by eliminating PAYG licenses.
Before optimization, Snow Atlas detects potential BYOL savings and on-premises licenses available for re-use.
This after view demonstrates actual savings and further optimization opportunities, as well as license compliance across the hybrid IT environment.
Additionally, Snow Atlas provides an easy view to understand your effective license position and where you may have a shortfall, so you can get compliant before an audit situation occurs. It is critical to proactively stay compliant as auditors often look three months into the past.
Since the original writing of this blog, the Flexera team was able to find an additional 1,024 Windows Server licenses that could be converted to BYOL, resulting in a total savings of nearly $500,000 a year. With cloud license optimization data at your fingertips, organizations can continue to see savings over time.
