How to reduce your Google Cloud Compute Engine Bill by 50% with Committed Use Discounts — Part 1

Committed Use Discounts are Low Effort and High Savings

Only about 20% of the Google Cloud users I speak with have a Committed Use Discount Strategy.

That translates to a lot of savings users are not taking advantage of. There are two versions of CUDs: resource-based, which are also known as standard CUDs, and spend-based, which are also known as flexible or flex CUDs. The basic strategy to maximize savings using CUDs is to use resource-based CUDs for planned steady-state workloads and spend-based CUDs for growth, bursty, or unplanned workloads.

Applying a mix of resource and spend based CUDs can potentially reduce your Google Cloud Compute Engine by half since Committed Use Discounts (CUDs) offer up to 57% savings on Google Cloud Compute Engine in exchange for a commitment of 3 years.

There are Two Types of Committed Use Discounts: Resource-Based and Spend-Based

Resource Based vs Spend Based (Flex) CUDs

Resource-Based CUDs or Standard CUDs:

• The discount only applies to a specific VM machine family within a specific region within a specific project.
• Up to 57% discount with a 3 year term. (Note the table above from a old blog post shows a discount of 55% but the documentation is updated with 57%).
• As the name suggests, resource based CUDs apply to Compute Engine resources such as vCPU and memory.

Spend-Based or Flexible (Flex) CUDs:

• The discount is applied at the billing account level so the discount can be shared across VM machine families, projects, and regions.
• Up to 46% discount with a 3 year term. Because of the flexibility to share the CUD across multiple VM machine families, projects, and regions, the trade-off is a slightly lower discount.
• As the name suggests, spend-based CUDs apply to a spend amount. For example, if you commit to spending $100/month on compute engine, then the CUD would apply to this $100 of spend.

A Committed Use Discount Strategy to Maximize Savings is to use both Resource-Based and Standard CUDs.

Apply a mix of Resource and Spend Based CUDs to maximize savings

1. For steady state workloads, such as production workloads, where you know what VMs you will use in what regions in what projects, maximize your discount by committing to a 3 year resource-based CUD.
2. For new and/or bursty workloads, such as a dev/test environment, where you’re unsure of what resources you will use in which regions and projects but you have confidence you will spend a certain amount, commit to a 1 or 3 year spend-based CUD to maximize your discount. After you experiment and understand what your future steady state looks like, then ‘convert’ some of your spend-based CUDs to a resource-based CUD to maximize savings.
3. For busty workloads that you cannot forecast, such as one-off experiments, then take advantage of sustained-use discounts (SUDs). SUDs are automatically applied to N1, N2, N2D, C2, M1, and M2 machine types after using a VM for at least 25% of a month so there is nothing you need to do to get this discount other than use the VM for an extended period of time.

Spend-based Committed Use Discounts are available for other Products

This blog post focuses on compute engine. However, spend-based committed use discounts are also available for other products such as AlloyDB for PostgreSQL, BigTable, CloudSQL, etc. The full list is available here. The spend-based CUD concept is the same for these products but the discounts will be different.

Ask about Commit and Enterprise Agreements to get even more Discounts

If you want to unlock even more discounts after you’ve maximized CUDs, then ask your Google Cloud Account Team about Cloud Commitments and Enterprise Agreements. These agreements are designed for customers who want to make a large commitment with Google across multiple products. Committed Use Discounts are self-serve in the Google Cloud Console while agreements would be negotiated with your account team. The general rule of thumb with agreements is the more money you commit to spending on Google Cloud and the longer you’re willing to commit to using Google Cloud, the larger a discount you can negotiate.

Summary

• CUDs can significantly reduce your Compute Engine bill, upwards of 50%.
• There are two types of CUDs: resource-based (standard) and spend-based (flex).
• A mix of resource-based and spend-based CUDs will maximize savings. Resource-based CUDs are ideal for steady-state workloads while spend-based CUDs are ideal for new or bursty workloads.
• After you’ve maximized savings using CUDs, talk to your Google Cloud account team about commit agreements for further savings.

Next Steps

In part 2 of this series, I will show how to determine, configure, and analyze your CUDs.

References

• https://cloud.google.com/compute/docs/instances/committed-use-discounts-overview
• https://www.googlecloudcommunity.com/gc/Community-Blogs/How-to-optimize-cloud-costs-with-Committed-Use-Discounts-for/ba-p/664178
• https://cloud.google.com/blog/products/compute/save-money-with-the-new-compute-engine-flexible-cuds
• https://cloud.google.com/compute/docs/instances/signing-up-committed-use-discounts

YouTube