Free Local AWS Emulators in 2026: What Each One Runs and Where Each One Stops
LocalStack's free edition ended in March 2026, and on 11 November AWS shipped its own emulator. Where each local option runs, what it costs and where it stops.
Two announcements that could cut your RDS bill by 35-55%. Here's what we learned and how we're applying it.
At INNOMIZE, we manage database workloads across multiple client projects. The stack is diverse: MySQL and PostgreSQL on RDS, DynamoDB for serverless applications, ElastiCache for caching layers. Instance types range from t3 and t4g for dev environments to larger instances for production.
We just finished auditing one project and found 12 RDS instances still running on db.t3, an instance family released in 2018. Not ancient, but not optimal either.
This audit came at the right time. At re:Invent 2025, AWS announced several database cost optimization features that change how we think about instance selection and commitment strategies. For us, this isn't just about cost savings. It's also about performance optimization: should we upgrade from t3 to t4g for better Graviton performance, or jump to r7g/r8g for more memory headroom?
This article covers the two biggest announcements: Database Savings Plans and new RDS for SQL Server/Oracle capabilities. We'll share our analysis and the migration path we're planning.
Database Savings Plans is a new flexible pricing model where you commit to a consistent $/hour usage over a 1-year term. No upfront payment required.
Notice the gap. Serverless gets 35%, provisioned gets 20%.
As we mentioned in our previous article about EKS Capabilities, AWS continues pushing teams toward managed and serverless offerings. The pricing tells the story: if you adopt Aurora Serverless v2, ElastiCache Serverless, or DocumentDB Serverless, you get nearly double the discount. AWS wants you on serverless. They're making it financially compelling.
Database Savings Plans apply to:
Unlike Reserved Instances, Database Savings Plans automatically apply regardless of:
This flexibility is significant. With Reserved Instances, you're locked to specific instance types. With Savings Plans, you commit to spend, not configuration.
Instance type restrictions:
Only latest-generation instances qualify for Database Savings Plans:
Older generations like t3, t4g, r6g, m6g are not eligible.
The t4g problem:
Here's an interesting situation. The db.t4g family, released in 2021 with Graviton2 processors, has no discount options at all. AWS promised Reserved Instance support back in 2021. Four years later, it's still "in works."
This creates a strange decision matrix:
For 24/7 dev environments, t3 with Reserved Instance can actually be cheaper than t4g on-demand. For variable workloads that spin up and down, t4g wins on hourly rate.
This is the exact situation we found in our audit.
Cannot stack discounts on the same workload:
From the AWS Database Savings Plan pricing page:
"For the same workload, you cannot combine Database Savings Plans with Amazon RDS Reserved Instances or Amazon DynamoDB reserved capacity discounts. However, you can buy a Reserved Instance or DynamoDB reserved capacity discount to apply to one workload and a Database Savings Plan to address another."
How it works in practice:
Reserved Instances take priority. If you have both RI and Savings Plan, AWS applies RI first, then Savings Plan covers remaining eligible usage:
Example: 3 db.r7g.large instances + 1 Reserved Instance + $50/hr Savings Plan commitment
You don't get 45% + 20% = 65% off on a single instance. Each discount applies to separate usage.
Strategic use: RI for predictable workloads on specific instance types, Savings Plan for flexible workloads where you might change engines, regions, or instance sizes.
For a db.r7g.large instance (us-east-1):
The tradeoff: Savings Plans offer less discount but more flexibility. Reserved Instances offer deeper discounts but lock you to specific configurations.
AWS also announced new capabilities for RDS SQL Server and Oracle that address cost optimization.
Amazon RDS for SQL Server now supports SQL Server 2022 Developer Edition.
Why this matters:
Previously, dev/test environments required Standard or Enterprise Edition licenses. Now you can use Developer Edition, which includes all Enterprise features with no licensing costs.
The savings:
For projects running multiple non-production SQL Server instances, this eliminates a significant cost.
Restrictions:
How to set up:
What changed:
New Intel-based M7i and R7i instances with a CPU optimization feature that can reduce costs by up to 55%.
How Optimize CPU works:
SQL Server licensing is per-vCPU. The Optimize CPU feature disables SMT (Simultaneous Multi-Threading), reducing vCPU count by 50% while maintaining the same physical cores.
Performance remains near-equivalent because you keep the same physical cores.
Instance improvements (M7i/R7i compared to M6i/R6i):
Reserved Instance support:
M7i and R7i instances support Reserved Instances with up to 46% savings compared to On-Demand.
RDS for Oracle and SQL Server now support up to 256 TiB storage per instance. Previously limited to 64 TiB.
How it works:
You can attach up to three additional storage volumes. This also provides 4x improvement in IOPS and I/O bandwidth.
Storage strategy:
You can mix io2 and gp3 volumes to optimize price-performance.
Back to our audit. We found 12 db.t3 instances across one client project: a mix of MySQL and PostgreSQL databases serving dev, staging, and production environments.
The t3 family served us well, but it's time to move. The question: migrate to t4g or jump to r7g/r8g?
Option A: t3 → t4g
Option B: t3 → r7g/r8g
The decision framework we're using:
The counterintuitive finding: for always-on dev environments, staying on t3 with Reserved Instance is cheaper than upgrading to t4g. The 37% RI discount on t3 beats the 10% on-demand savings of t4g.
For production, we're moving to r7g. The memory increase alone justifies it, and qualifying for Database Savings Plans or Reserved Instances makes it financially sound.
The SQL Server Developer Edition announcement changes our approach for .NET projects:
Before:
After:
Expected savings:
Example calculation:
One project runs 3 non-production SQL Server instances on db.r5.large (2 vCPU, 16 GB):
For larger instances like db.r5.xlarge or db.r5.2xlarge commonly used in staging environments, the license portion is even higher. Across multiple non-production environments, Developer Edition can easily save $1,500-2,500/month depending on instance sizes.
Actions:
These re:Invent announcements aren't just press releases. They're pricing signals.
AWS is telling us: move to serverless (35% discount), use latest-gen instances (Savings Plan eligible), and stop paying for SQL Server licenses in dev environments (Developer Edition).
For our t3 fleet, the path is clear: