Can Serverless Really Reduce Your Cloud Costs?

Exploring the use cases when serverless can save you money and when best to use it.

September 29, 2024

Read on Beehiiv | Sep 29, 2024

Welcome to The Cloud Economist!

Last week I talked about the most popular cloud service, Amazon S3, its different storage classes, and how understanding and using these can help you optimize significant costs.

I shared some interesting articles and tips on how S3 can be used for a low-cost storage solution for almost any need.

If you missed that you can read it here.

Last Week’s Question

In the question of the week, I asked how can you optimize costs when uploading multiple files to Amazon S3.

Amazon S3 API costs depend on the type of call, not file size, so uploading small or large files costs the same per request.

However, uploading many small files can add up quickly. To save on costs, you can batch small files into one using tar and upload them in bulk, reducing the number of API calls.

This week let’s discuss a more controversial subject: how serverless can reduce your cloud costs.

Spoiler: it’s all about your use case.

Here are the best articles I’ve found on cloud cost savings this week, summarized.

The article shares these 7 cost-optimization strategies:

  1. Pick the Right Service Provider: 

    • Choose a cloud provider (AWS, Azure, Google Cloud) that fits your needs and pricing model.

    • Consider alternatives like PlanetScale or Firebase for specific tasks.

  2. Monitor Resource Usage:

    • Track resource usage with monitoring tools to optimize performance and set usage alerts.

  3. Optimize Code and Functions

    • Write efficient, lean code to reduce execution time, minimize dependencies, and use caching.

    • Set appropriate memory limits and function timeouts.

  4. Implement auto-scaling

    • Set auto-scaling rules to adjust resources based on workload demands, with caution on upper limits.

  5. Minimize Compute usage

    • Utilize services that reduce the need for compute, like AWS Step Functions to avoid unnecessary Lambda invocations.

  6. Track costs KPIs

    • Monitor KPIs to ensure serverless costs align with customer growth.

AWS’s offers 5 design principles for cost optimization.

These include:

  • implement cloud financial management to enhance financial success.

  • adopte a consumption model to pay only for what you use.

  • measure overall efficiency to balance output and costs.

  • AWS also recommends avoiding heavy lifting by leveraging managed services

  • analyze expenditure to attribute costs to workload owners, helping optimize resources and improve ROI.

These principles guide businesses in managing cloud costs effectively while focusing on core objectives.

Read the article above for a deeper dive into these strategies.

One Tip on Cloud Cost Savings

Serverless is quite controversial when it comes to costs. The best strategy here is to first understand your use case and needs.

Not every application requires serverless. Here is a general guide on when serverless will actually reduce your costs:

  • ephemeral transactions (like in social media use cases)

  • applications that don’t have heavy analytics processing

  • Smaller to medium-sized applications

  • Hybrid workloads that fit the serverless model (such as sending image processing to serverless compute but using servers to handle storing the image)

  • Prototyping and proof of concepts.

I write often about when and when not to use serverless, I recommend my blog to learn all about this.

This Week’s Question

How can you minimize the impact of cold starts in a serverless function to reduce costs?

Until next week.

The Cloud Economist