Understanding License Recommendations
The license recommendations tab is an AI/ML based feature that delivers actionable, data-driven insights to help your organization optimize licensing costs for Microsoft SQL Server and Windows Server for instances running in public cloud environments such as AWS and Azure.
The feature combines rule-based logic with an AI/ML-driven analysis to recommend the most cost-efficient licensing model, either Bring Your Own License (BYOL), or Pay As You Go (PAYG). There are several key factors for these recommendations:
- Resource usage over time
- Assigned and available BYOL entitlements
- Cost for each cloud license model based on utilization across cloud environments
The license recommendations tab provides further opportunities for data-informed licensing optimization by identifying underutilized BYOL entitlements, quantifying underutilization and associated cost waste, and providing actionable insights into relevant resources in the cloud business account.
The license recommendations tab can provide the following insights:
Detect underutilized BYOL licenses—Identify cloud resources (instances or VMs) that are running SQL Server or Windows Server, and are assigned BYOL entitlements but are not utilizing them fully due to vendor minimum vCPU license compliance requirements. Quantify the percentage of underutilized BYOL entitlements and the associated cost waste, and provide visibility into which business account resources are not fully utilizing BYOL entitlements.
Recommend PAYG to BYOL conversion—Recommend conversion of high-spend PAYG resources to BYOL entitlements in cases where spare BYOL entitlements are available. Recommendations are prioritized based on projected cost savings, resource size (vCPU) and usage patterns in each cloud. The recommendations to convert PAYG resources to BYOL licenses are available when the projected savings are higher than 10%.
Recommend BYOL to PAYG conversion—Recommend conversion of BYOL entitlements to a PAYG model in cases where the PAYG cost based on the actual usage is lower than the BYOL license cost for the whole year. These recommendations consider the availability of BYOL licenses, as well as usage patterns which suggest a PAYG model would be more cost effective.
Resource Eligibility and Filtering—Recommendations exclude any non-production resources (development or testing), leveraging pre-calculated eligibility flags from CLM (i.e. development or testing). Factors considered include cloud, vCPU usage, region and business accounts for precise targeting, allowing the relevant teams to take appropriate action to run software in a more cost-efficient way.
Recommendations use the latest BYOL unit cost from IT Asset Management (ITAM), and in case the unit cost is not available, the recommendations will fall back to the average BYOL unit cost, as per the table below:
| Product | Average BYOL Unit Cost |
|---|---|
| SQL Server Standard Edition | $700 |
| SQL Server Enterprise Edition | $2,800 |
| Windows Server Standard Edition | $60 |
| Windows Server Datacenter Edition | $240 |
Within the tab, two tables provide a detailed breakdown of instances that can be converted from BYOL to PAYG, and vice versa, from PAYG to BYOL. The following information is available in the tables:
- Instance Name —The name of the instance.
- Cloud Provider—The cloud provider hosting this cloud resource.
- Account Name—The name of the customer account with the cloud provider for this cloud resource.
- Name—The product and edition name of the software this recommendation is for.
- vCPU—Number of vCPUs the instance currently uses.
- Potential Savings—Total amount of projected savings in the next 365 days if recommendation is applied.
- Projected PAYG Spend—Projected PAYG spend for the next 365 days.
- Projected BYOL Spend—Projected BYOL spend for the next 365 days.
- Projected Annual Usage Hours—Projected usage hours for the next 365 days.
- Minimum Hours To Breakeven—Number of hours the instance should be active to break even.
- ResourceId—The unique resource ID for this cloud resource.
The tab also provides a table that details underutilized BYOL entitlements. In addition to the essential details for each instance (Instance Name, Cloud Provider, Account Name, vCPU, Name, ResourceId), the following columns are available:
- License Required—Number of licenses the instance requires with its current license model.
- License Model—The current license model the instance uses, can be BYOL or PAYG.
- Usage Hours—Total number of hours the instance has been active.
- Underutilized BYOL Cost—Total amount of cost incurred by underutilized BYOL entitlements.
- BYOL Underutilization %—Percentage of underutilized BYOL entitlements.
For each recommendation, you can access the recommendation details page by right-clicking the instance, and clicking Drill through > Recommendation Details. The recommendation details page contains the essential details for each instance (Instance Name, Cloud Provider, Account Name, vCPU, Name, License Model, Resource ID), as well as the following columns:
- Instance Type—The type of instance from the range of types the cloud provider offers.
- BYOL Eligible—Whether the instance can be converted to a BYOL license model.
- PAYG Price/hr—The price per hour the instance would incur with a PAYG license model.
The recommendation details page also displays a usage chart, showing the hours the instance has utilized until the current moment, as well as a prediction of hours the instance is projected to use in the future.
The AI/ML recommendation engine is designed by Flexera to operate within the Flexera One platform, applying intelligence to customer software usage to identify opportunities for software cost reduction and license optimization. The engine generates customer-specific outcomes and does not use customer data for training or sharing.