Service eligibility and regional restrictions
PuppyIP serves only compliant overseas businesses and their authorized personnel. Proxy services are not available in mainland China. The service may only be used for lawful business activities outside mainland China. Use of this service within mainland China is prohibited.
Hosting a proxy IP or server overseas does not change these restrictions. The service must not be provided to end users in mainland China through relaying, forwarding, sharing or resale. Before use, read the Terms of Service.
Key Takeaways
- Use separate Aurora and RDS support tables; compatibility with the same database does not establish identical instance-class requirements.
- Explicit Aurora version ranges are not a promise that every later version is supported.
- Price-performance is different from a bill discount. Verify applicable rates, deployment and storage costs separately.
Start with the service and Region
AWS announced Graviton5-based R9g general availability on November 12, 2026 for Aurora PostgreSQL-compatible and MySQL-compatible editions, plus RDS for PostgreSQL, MySQL and MariaDB. The four announced Regions are US East (N. Virginia and Ohio), US West (Oregon), and Europe (Frankfurt).
This evaluation concerns the managed database class db.r9g. Record the actual service, Region, full engine version and target size before opening the support table. A processor name in another service, an option in another Region, or an inventory entry that only says “MySQL” is insufficient evidence.
Read Aurora and RDS version rules separately
For example, the current Aurora PostgreSQL db.r9g.large row lists 17.4 through 17.7 and 17.9 through 17.11. Those ranges do not include 17.8. Do not rewrite them as “17.4 and every later version.” Match the full version and consult the corresponding row for other major versions and sizes.
The RDS row for that size instead lists MySQL 8.4.5 and higher, and PostgreSQL minimums within its listed branches: 14.18, 15.13, 16.9, 17.5 and 18.1. Do not apply these thresholds to Aurora or extend them to unlisted major versions. Recheck the current documentation and available options in the target Region before acting; these examples are not a complete compatibility list.
Keep the performance claim within its stated scope
AWS reports up to 30% better performance and up to 22% better price-performance for Aurora against equivalent-size Graviton4 instances, using on-demand pricing for the latter comparison. Results depend on engine, version and workload. These figures do not apply automatically to every RDS engine or every database.
Price-performance compares price with work completed; the bill also depends on runtime, sizing and other charges. Define the business goal first—for example, query latency and cost per unit of work under the same load. The announcement’s upper bound is not a measurement of your workload or a guaranteed 22% bill reduction.
Include more than the instance price
Aurora pricing distinguishes Standard from I/O-Optimized configurations, with different instance, storage and I/O charging models. Include primary and replica instance charges. For RDS, open the pricing page for the actual engine. The captured public pricing text did not establish an exact R9g hourly rate, so that price remains unknown; do not enter zero or derive it from the price-performance percentage.
Build a comparison with the same Region, service and engine, instance size, deployment, purchasing option, runtime, storage, applicable I/O and data transfer. Hold these assumptions constant before listing proposed changes. Identify cost differences from a Region or configuration change separately. An on-demand comparison does not establish the result for a reserved purchasing option.
Create a reviewable evaluation before scheduling changes
First retain the existing configuration and workload baseline. Next check the target service, Region, version and size. Then estimate complete costs using one set of usage assumptions. Within an authorized test scope, compare throughput, latency, error rate and cost, recording the sample period and load conditions. No benchmark or database change was performed for this article.
Class support alone does not establish that a particular existing instance can switch directly or without downtime. The migration path and downtime for your deployment have not been verified here. If implementation is warranted, consult the engine-specific procedure, maintenance impact and recovery plan, then follow the team’s normal change approval process.
Sources
Frequently Asked Questions
Can RDS for Oracle, SQL Server or Db2 use R9g too?
The captured RDS support table marks all three as No for db.r9g.large. Membership in RDS does not establish support; check the actual size and current engine table.
Will a memory-optimized instance fix slow queries?
An instance type alone does not identify the bottleneck. Use workload evidence to distinguish resource, query-plan and storage effects, then choose evaluation metrics. A hardware announcement does not diagnose an individual query.