How do I estimate my AWS S3 costs?

If you’ve ever tried navigating AWS documentation, you know it can feel like diving into a sea of jargon, especially when it comes to understanding S3 pricing. You’re not alone! With so many storage classes and cost structures, it’s easy to feel overwhelmed. But pricing isn’t as complicated as it seems.

Once you break it down, it becomes clear how to save money while meeting your storage needs. Let’s simplify the chaos and explore S3’s storage classes, when Intelligent-Tiering doesn’t make sense, and the alternatives you should consider.

AWS S3 Storage classes and pricing:

Amazon S3 (Simple Storage Service) is one of the most popular storage solutions on AWS, offering a range of storage classes designed to optimize costs based on data access patterns.

As of writing of this article, when you go to the official AWS pricing page: https://aws.amazon.com/s3/pricing/?nc=sn&loc=4, one sees more than 9 storage categories.

Some of the major categories include

• S3 Standard

• Standard-IA

• One Zone-IA

• Glacier Instant Retrieval

• Glacier Flexible Retrieval

• Glacier Deep Archive

• Intelligent Tiering

• and more.

But do you really need so many of these for your project?

Here’s a quick breakdown of four main S3’s key storage classes that you need to know and their use case:

1. S3 Standard: Designed for frequently accessed data, with low latency and high durability. Ideal for primary storage and active content.

   • Pricing: ~$0.023/GB per month in most regions.

2. S3 Standard-IA (Infrequent Access): For less frequently accessed data that requires rapid retrieval.

   • Pricing: ~$0.0125/GB per month, plus retrieval fees.

3. S3 Glacier Instant Retrieval: Optimized for archival data that needs immediate access.

   • Pricing: ~$0.004/GB per month.

4. S3 Glacier Deep Archive: The lowest-cost option for data that is rarely accessed, with retrieval times of up to 12 hours.

   • Pricing: ~$0.00099/GB per month.

What I personally found confusing was the category called S3 Intelligent-Tiering and how does it relate to the above ones.

S3 Intelligent-Tiering in itself is an own category, with its further sub-categories.

When you go to official documentation here is another table for intelligent tiering pricing:

S3 Intelligent-Tiering automatically moves data to the most cost-effective tier based on access patterns. Suitable for unpredictable workloads.

• Pricing: ~$0.023/GB per month for the Frequent Access tier, with small monitoring fees.

In order for intelligent tiering to work, object size must be more than 128 Kb else they are not monitored. And there is a monitoring fee of 0.25 cents per 1000 objects or to make it simple, 2.5 cents per 10,000 objects.

When Intelligent tiering does not make sense?

While one would think that S3 Intelligent-Tiering is a no brainer, there are scenarios where it might not be the best fit:

1. Static Access Patterns: If you already know how often your data will be accessed, using a fixed tier/or using the lifecycle rules in s3 bucket to transition to low fee teirs like S3 Standard-IA or S3 Glacier can save you cost by avoiding monitoring fees.

2. Small Datasets: For smaller datasets, the monitoring and automation fees of Intelligent-Tiering can outweigh the savings.

3. Archive Data: If your data is strictly archival with no need for frequent access, S3 Glacier Deep Archive is more cost-effective.

Besides the storage, S3 also charges for:

Outbound Data Transfers

• Costs for data transferred out of S3 to the internet or other regions.

• Free for data transferred into S3.

• Free for data transferred to AWS services in the same region (e.g., EC2).

API Requests and Operations

S3 will also charge for API requests such as Listing buckets, reading from buckets.

• operations include PUT, COPY, POST, LIST, or GET requests made to S3 either via console or via API.

• Example: S3 Standard-IA charges higher for retrievals than Standard.

This is a brief summary of AWS S3 pricing. You can visit https://calculator.aws/#/createCalculator/S3 to create an estimate of you application.

Conclusion and summary:

If you want to anticipate how much S3 is going to cost you, have an idea of 3 things:

1. How much storage you need?

2. How much outbound data transfer do you expect?

3. How many API calls and operations do you anticipate?

Because AWS is going to invoice you for all of these actions. Once you have an estimate for these questions you can use AWS pricing calculator to find the exact cost in your region.