AWS has introduced a game-changing feature for organizations running Oracle databases on Amazon RDS: extended memory configurations for RDS compute instances using the BYOL (Bring Your Own License) model. These new configurations enable significant cost savings by optimizing Oracle licensing requirements, providing a vital tool for organizations looking to reduce their IT expenses.

The Oracle Licensing Landscape in the Cloud

Licensing Oracle software in the cloud is not as straightforward as on-premises environments. Under Oracle’s standard licensing agreements (e.g., Oracle Master Agreement), there is no explicit right to license Oracle products based on vCPUs. However, Oracle’s Cloud Compute Policy, a non-contractual document - offers guidelines for licensing Oracle on approved cloud platforms like AWS and Azure. 

This policy allows organizations to:

• Count two vCPUs as equivalent to one Oracle Processor license when hyper-threading is enabled (the default in AWS).

• Count one vCPU as equivalent to one Oracle Processor license when hyper-threading is disabled.

While this policy provides a framework for cloud licensing, it often results in higher licensing costs compared to on-premises deployments due to the inherent differences in how resources are measured.

Introducing Extended Memory Configurations in Amazon RDS for Oracle

AWS offers various instance types for RDS for Oracle, each with different memory-to-vCPU ratios. The r5 family, known for its memory optimization, has been a popular choice due to its 8:1 memory-to-vCPU ratio. However, for workloads requiring higher memory without additional vCPU overhead, the traditional r5 configurations could lead to unnecessary licensing costs.

AWS addressed this challenge by introducing extended memory configurations for RDS for Oracle. These new instance sizes allow users to provision significantly higher memory without increasing the number of vCPUs. For example:

• Traditional r5 instances offer up to an 8:1 memory-to-vCPU ratio.

• Extended memory configurations provide ratios as high as 64:1, enabling customers to allocate the memory their databases need without inflating their Oracle licensing liability.

Real-World Impact: Cost Savings Through Optimization

Consider a scenario where you need 12 vCPUs and 192 GiB of memory for your Oracle database. With traditional r5 instances, you’d need to choose a db.r5.8xlarge instance, doubling your vCPUs and Oracle licensing liability compared to a db.r5.4xlarge. By leveraging the extended memory configurations, you can achieve the same memory requirements without increasing your Oracle Processor licenses, translating to substantial cost savings.

For context, the list price of a single Oracle Database Enterprise Edition Processor license is $47,500. By reducing your licensing needs, the incremental cost of extended memory instances becomes negligible compared to the potential savings on Oracle licensing.

Why It Matters

This feature is exclusive to Amazon RDS for Oracle and highlights AWS’s commitment to addressing the unique challenges of running Oracle databases in the cloud. While the extended memory configurations may increase your AWS spend, the reduction in Oracle licensing costs often far outweighs the additional expense, offering a compelling value proposition.

Simplify Your Oracle Licensing Journey

Navigating the complexities of Oracle licensing in the cloud can be daunting, but you don’t have to do it alone. At Cintra, we specialize in helping organizations optimize their Oracle environments to achieve the best possible cost efficiency. Whether you’re already running Oracle on AWS or planning a migration, our experts can guide you in leveraging features like extended memory configurations to maximize your savings.

Contact us today to learn how we can help you transform your Oracle licensing strategy and unlock new levels of efficiency in the cloud.