The evolution of cloud infrastructure offers numerous opportunities for companies; however, the increase in cloud expenditure is a major pain point for many. For AWS customers, choosing the right database can greatly impact their budget management while planning for future expansions.
Amazon RDS is still the most trusted and popular managed relational database solution on AWS with over 2 million users across the globe as well as supporting a wide array of workloads and database engines. AWS Aurora Serverless, popular for auto-scaling and fully managed features, works best with applications that have unpredictable or highly variable workloads. RDS works best with predictable workloads because it is lower cost and more reliable in the long term.
In this blog, we will walk you through the pricing breakdown for Aurora Serverless and RDS, help you identify which service aligns with your requirements, and guide you through the steps. By the end, you will gain valuable insights into reducing cloud costs and optimizing your configuration.
What Are Aurora Serverless and AWS RDS?
Before we compare the two, it is important to analyze each service individually, along with its pricing structure.
What Is Aurora Serverless?
Amazon Aurora Serverless is a managed database service offering where both compute and memory resources scale automatically based on the needs of the workload. This type of database does not require the user to provision a specific amount of capacity beforehand. Rather, payment is based on actual usage.
Key Features:
Per-Second ACU Billing: You are charged for each ACU consumed, which scales with demand.
Auto-Scaling: The service automatically increases or decreases capacity for traffic bursts or drops.
Ideal For: Applications with low or spiky workloads, unpredictable traffic, or test development environments.
What Is AWS RDS?
AWS RDS is a fully managed cloud database service, allows you to easily set up, operate, and scale databases like MySQL, PostgreSQL, MariaDB, Oracle, and SQL Server. With RDS, database management operations like backups, patching, and failover are done automatically so that you can focus on building and optimizing applications instead of managing databases.
Key Features:
Instance-Based Pricing: Pricing is based on the compute instance size/type and storage used.
Reserved Instances: Offers discounts up to 75% for commitments of long-term usage.
Ideal For: Applications with steady or predictable traffic, sustained workloads, and budget-friendly companies.
How Aurora Serverless Costs Add Up
The pricing model of Aurora Serverless can be problematic for workloads with steady or predictable usage patterns, despite offering flexibility for variable workloads.
Understanding ACU Billing
Aurora Serverless charges per second based on ACUs, and each ACU contains 2 GB of RAM, compute, and networking. Costs are prorated/rounded to the nearest hour, with a minimum capacity of 0.5 ACUs, costing $0.06/hour when idle.
Example Costs:
Scenario 1: Steady workload at 8 ACUs for 6 hours/day (180 hours/month) at $0.12 per ACU-hour = $172.80/month.
Scenario 2: 24/7 workload at 8 ACUs (730 hours/month) = $700.80/month, which is 4x more expensive than an RDS r5.large instance ($175.20/month).
Hidden Costs in Storage and I/O
Beyond ACUs, Aurora Serverless also charges for storage and I/O operations. While each single read or write has little impact, the ensuing costs for high I/O applications can lead to shocking bills.
When Aurora Serverless Becomes Less Affordable
For workloads with steady usage or predictable traffic patterns, the automatic scaling feature of Aurora Serverless becomes a needless expense. Companies often find that the flexibility per-second ACU billing provides is not worth the premium charged as compared to RDS’s fixed-instance pricing.
Why RDS is More Cost-Effective for Many Workloads
AWS RDS can prove exceptionally useful for companies planning for long-term workloads as it offers predictable pricing as compared to other database services, making AWS RDS advantageous in terms of cost savings.
Steady vs. Spiky Workloads
Steady Workloads: RDS is the optimal choice for workloads with a steady level of demand. You can resize your instance to the optimal level and avoid ancillary costs related to scaling or surplus capacity, as is the case with Aurora Serverless. Budgeting is simpler when there is fixed pricing on an instance which RDS offers.
Spiky Workloads: RDS can also meet variability using burstable instance types like t3.micro, or by planning for peak capacity, with Aurora Serverless excelling in instances of unpredictable traffic spikes. For all but the most extreme production workload spikes, RDS remains the cheaper solution.
Benefits of Reserved Instances
By committing to a 1 or 3-year Reserved Instance, you can save up to 69% compared to On-Demand pricing. With flexible payment options (All Upfront, Partial Upfront, or No Upfront), it's easier to balance savings with cash flow.
Example Savings:
Aurora Serverless (8 ACUs, steady workload): $700.80/month or $8,409.60/year (compute only)
RDS Reserved Instance (r5.large, 1-Year No Upfront): $94.17/month or $1,130/year
Annual Savings: Depending on your region and setup, annual savings could exceed 86%.
RDS gp3 Storage Advantages
Free IOPS: RDS gp3 comes with 3,000 IOPS that incur no additional cost and only requires payment for additional IOPS if they are provisioned. This is different from Aurora, which charges separately for I/O requests, making RDS more cost-efficient for high I/O workloads.
Flexible Scaling: With gp3, there is the ability to scale both the storage and IOPS separately, which optimizes both performance and costs.
Aurora and RDS Savings Example
Example 1: Steady Workload (8 ACUs)
Scenario: A database requiring 16 GB RAM (8 ACUs in Aurora vs. an r5.large instance in RDS).
Aurora Serverless Cost:
Compute: 8 ACUs × $0.12/hour × 730 hours/month = $700.80
Storage (100 GB): 100 GB × $0.20/GB-month = $20
I/O (10 million requests): 100 million × $0.20 = $20
Total: $700.80 + $20 + $2 = $722.80
AWS RDS Cost (Reserved Instance):
Compute (db.r5.large): $0.17/hour × 730 hours = $124.10
Storage (100 GB gp3): 100 GB × $0.10 = $10
I/O (5,000): 3,000 is free + 2,000 × $0.005 = $10
Total: $124.10 + $10 + $10 = $144.10
Savings:
($722.80 − $185.20) ÷ $722.80 × 100 = 80.1%
Example 2: High I/O Workload
Scenario: A database requiring 1TB storage (8 ACUs and 5000 IOPS in Aurora vs. an r5.large instance in RDS).
Aurora Serverless Costs:
Compute: 8 ACUs × $0.156 per ACU-hour × 730 hours = $911.04
Storage: 1,000 GB × $0.225 per GB-month = $225
I/O: Included (no separate IOPS charge in I/O-Optimized mode)
Total: $911.04 (compute) + $225 (storage) = $1,136.04
AWS RDS (r5.large, io3 storage, provisioned IOPS) Costs:
Compute: $0.24 per hour × 730 hours = $175.20
Storage: 1,000 GB × $0.125 per GB-month = $125
IOPS: 5,000 IOPS × $0.10 per IOPS-month = $500
Total: $175.20 (compute) + $125 (storage) + $500 (IOPS) = $800.20
Savings:
($1,136.04 − $800.20) ÷ $1,136.04 × 100 = 29.56%
How to Decide If Migration Is Right for You
Evaluate your workloads and costs prior to switching to determine if RDS is a better option for you.
Signs You Should Migrate
High ACU Usage: If your compute costs will likely be higher than those of an RDS instance, your Aurora Serverless cluster runs frequently at high ACU levels.
Steady or Predictable Traffic: Because RDS has a fixed pricing model, it can often be subsidized with Reserved Instances. Thus, it is more economical than Aurora Serverless, which is built to accommodate spiky workloads.
High Storage or I/O Demand: Aurora Serverless charges separately for storage and I/O operations, which can add up for high-demand workloads. RDS’s gp3 storage includes 3,000 free IOPS per GB, making it more budget-friendly for high I/O needs.
Interest in Reserved Instances: RDS offers Reserved Instances, saving up to 69% on costs for 1 or 3-year usages, an option not available with Aurora Serverless.
Using AWS Native Tools
Leverage AWS tools to analyze your usage and costs:
AWS Pricing Calculator: Estimate whether the costs of both Aurora Serverless and RDS meet your specific workload and region.
AWS Cost Explorer: Review historical ACU usage, I/O patterns, and storage costs to identify the use of those resources over time and estimate future costs.
Other Tools: Use AWS Trusted Advisor and AWS Budgets to monitor and control costs effectively.
Is Migrating Right for You?
For companies looking to enhance cost management and long-term performance, RDS is a strong contender when compared to Aurora Serverless. While there are scenarios where one might need Aurora Serverless, RDS does far better in scenarios of consistent and predictable workloads.
Review your expenses, plan them using the AWS Pricing Calculator, and think about migrating if the numbers add up to your expectations.
Join Pump for Free
If you are an early-stage startup that wants to save on cloud costs, use this opportunity. If you are a start-up business owner who wants to cut down the cost of using the cloud, then this is your chance. Pump helps you save up to 60% in cloud costs, and the best thing about it is that it is absolutely free!
Pump provides personalized solutions that allow you to effectively manage and optimize your Azure, GCP and AWS spending. Take complete control over your cloud expenses and ensure that you get the most from what you have invested. Who would pay more when we can save better?
Are you ready to take control of your cloud expenses?
