Optimizing Cost Allocation in SAP BTP: Shared Instances

[Eran]

In the evolving landscape of FinOps within SAP BTP, understanding shared instances becomes crucial. Many companies run a single instance of an SAP BTP service, such as the Integration Suite & HANA Cloud, while multiple departments use it for different purposes. However, the challenge arises when trying to allocate the costs of this shared resource fairly across departments. This post explores strategies for managing such scenarios efficiently.

Understanding the Cost Allocation Challenge

SAP services often charge based on a combination of a flat subscription fee and/or variable consumption-based pricing. For example:

• SAP Integration Suite: incurs a base fee for the instance, plus an additional cost based on the number of messages processed.
• SAP HANA Cloud: charges based on allocated storage and computing resources.
• SAP AI Core: costs depend on the number of large language models (LLMs) used and the compute resources consumed.

Potential Solutions for Cost Allocation in Shared Instances

To improve cost allocation within the SAP Integration Suite, SAP is exploring incremental enhancements to provide better tracking and transparency for shared instances. The approach focuses on:

• Enhancing Package-Level Information: Future updates may include better visibility into package-level consumption, helping organizations estimate and distribute costs more effectively.
• Aggregated Message Summaries: Rather than daily message counts, improvements in reporting could allow customers to track messages per month at the integration flow or package level.
• Improved Data Availability for Cost Allocation: SAP aims to make relevant consumption data more accessible, reducing the need for manual aggregation and facilitating more accurate cost distribution.

As these improvements take shape, feedback from users will be crucial in refining the approach and ensuring it meets real-world business needs.

The Workaround: Downloading and Analyzing Usage Data

Since fine-grained metering is unavailable, customers can manually extract daily usage data from the SAP Integration Suite cockpit. By aggregating this data monthly and associating it with specific integration flows, businesses can estimate departmental consumption and manually distribute costs.

This process involves:

• Exporting daily/monthly message count data from SAP BTP.
• Assigning message volume per department based on predefined rules.
• Allocating the shared instance's cost proportionally.

Addressing Cost Allocation Challenges in SAP HANA Cloud

One of the key challenges in cost allocation for SAP HANA Cloud is that current tracking methods primarily focus on data persistency (disk and memory usage) rather than compute consumption (CPU, calculations, and network egress). This can lead to an incomplete representation of actual resource utilization, particularly for workloads that are compute-intensive.

To enhance cost tracking accuracy, an alternative approach is being explored using stored procedures that can distribute costs within an instance more effectively. This method demonstrates how costs can be allocated at a more granular level based on schema or HDI container usage. However, for this to work, customers would need to deploy the stored procedure at the database administration level and grant the necessary access for execution.

While this approach introduces additional configuration steps, it offers a potentially viable method for organizations looking for a more refined way to allocate HANA Cloud costs among business units. Future improvements may focus on integrating these capabilities into a centralized, automated FinOps framework that can provide a more seamless cost allocation experience across multiple SAP BTP services.

Shared Instances in the FinOps Application

In addition to the strategies discussed for SAP BTP, the FinOps application provides a built-in capability to cross-charge shared instances using Managed Tags. This feature allows organizations to allocate costs for shared instances based on predefined percentage splits rather than direct consumption tracking.

How Managed Tags Work

With Managed Tags, businesses can assign cost allocations to different departments based on fixed percentages. For example:

• A HANA Cloud or Integration Suite instance can be tagged so that 20% of the cost is allocated to Department A and 80% to Department B.
• This approach provides a structured way to distribute costs for productive instances that have stable, predictable usage patterns.

While this method is not as dynamic as real-time metering based on actual consumption, it offers a practical workaround for companies where usage patterns remain consistent over time. This ensures fair cost distribution without the need for daily or monthly usage tracking.

By leveraging Managed Tags within the FinOps application, organizations can establish a systematic and efficient way to handle shared instance cost allocation, particularly for environments with static resource consumption.

Call to Action

We invite our readers to share their experiences and insights:

• Have you implemented cost allocation strategies for shared instances in your organization?
• Are there other SAP BTP services where you've encountered similar challenges?
• Do you have unique use cases or workarounds that could benefit the community?

For more information on the services discussed, explore the following resources:

• SAP Integration Suite: SAP Integration Suite - SAP Help Portal
• SAP HANA Cloud: SAP HANA Cloud - SAP Discovery Center
• SAP AI Core: SAP AI Core - SAP Discovery Center

Conclusion

Managing shared instance costs within SAP BTP requires a combination of manual tracking, process adjustments, and upcoming SAP improvements. Until granular metering is available, companies can rely on workarounds and SAP's planned reporting enhancements to optimize cost allocation.