gcpcommitted-use-discountscompute-enginecost-optimization

GCP flexible committed use discounts cost: spend-based commitments explained

Google Cloud offers two kinds of committed use discount, resource-based and spend-based flexible commitments, and the flexible kind trades some discount depth for the freedom to change machine types and regions. Here is how they compare on cost.

The C3X Team··6 min read

Quick answer

Google Cloud has two committed use discount models. Resource-based commitments lock in a specific amount of vCPU and memory in a region for one or three years at the deepest discount, often up to 55 percent or more for three years, but they only apply to that machine family and region. Flexible (spend-based) commitments instead commit to a steady hourly dollar spend on Compute Engine across families and regions, at a somewhat smaller discount, around 28 percent for one year and 46 percent for three years, in exchange for the freedom to change machine types and regions. The rule is use resource-based commitments for a stable, known machine footprint to maximize the discount, and flexible commitments when your machine mix or regions will change and the flexibility is worth a few points of discount.

Committed use discounts are Google Cloud's main lever for reducing steady compute cost, and there are two flavors that differ in both discount depth and flexibility. Resource-based commitments give the deepest discount but pin you to specific resources, while flexible spend-based commitments give a slightly smaller discount in exchange for the freedom to change what you run. Picking the right one depends on how stable your machine footprint is.

The two commitment types

AspectResource-basedFlexible (spend-based)
Commit toSpecific vCPU and memory, region, familyHourly dollar spend on compute
Approx. one-year discountAbout 37 percentAbout 28 percent
Approx. three-year discountUp to 55 percent or moreAbout 46 percent
FlexibilityLocked to family and regionAny family and region

A resource-based commitment locks in a specific quantity of vCPU and memory for a particular machine family in a particular region, for one or three years, at the deepest discount, roughly 37 percent for one year and up to 55 percent or more for three years on many families. A flexible spend-based commitment instead commits to a steady hourly dollar amount of Compute Engine spend that can be spread across any machine family and region, at a somewhat smaller discount, around 28 percent for one year and 46 percent for three years.

What the flexibility buys

The flexible commitment's value is that it does not care what machines you run. If you commit to a dollar-per-hour spend and later migrate from N2 to a newer family, or shift workloads between regions, the commitment keeps applying, because it is tied to spend rather than to specific resources. A resource-based commitment, by contrast, would strand if you moved off the committed family or region, forcing you to buy a new commitment or lose the discount. So flexibility protects you against the discount becoming useless when your architecture evolves.

What the flexibility costs

That protection costs a few points of discount. For a stable footprint you are confident will not change, the resource-based commitment's deeper discount is simply more savings on the same usage, so leaving those points on the table has a real cost. The flexible commitment only pays off when the flexibility is actually used or genuinely likely, if you commit flexibly but never change families or regions, you paid a discount premium for an option you did not exercise. This mirrors the general committed use discount tradeoff.

Choosing the right commitment

Assess how stable your machine footprint is over the commitment term. If you run a known, steady family in a fixed region and expect that to hold, take resource-based commitments for the deepest discount. If your machine mix, families, or regions are likely to change, perhaps because you are modernizing or growing across regions, the flexible spend-based commitment protects the discount at a modest cost. Combine either with sustained use discounts on usage above the commitment. Price your committed footprint against the resource catalog so the commitment type is chosen on real numbers before you buy.

FAQ

What is the difference between resource-based and flexible committed use discounts?

Resource-based commitments lock in a specific amount of vCPU and memory for a particular machine family and region, at the deepest discount, roughly 37 percent for one year and up to 55 percent or more for three years. Flexible spend-based commitments commit to a steady hourly dollar spend on compute across any family and region, at a smaller discount, around 28 percent for one year and 46 percent for three years.

When should I use a flexible committed use discount?

When your machine mix, families, or regions are likely to change over the commitment term, perhaps because you are modernizing to newer machine families or growing across regions. The flexible commitment stays applied as you shift workloads because it is tied to spend rather than specific resources, protecting the discount from becoming stranded, which is worth the few points of smaller discount.

Why do flexible commitments give a smaller discount?

Because you are paying for flexibility. A flexible spend-based commitment applies to any machine family and region, so Google offers a somewhat smaller discount than a resource-based commitment that pins you to specific resources. The flexible commitment only pays off when the flexibility is used or genuinely likely; if you never change families or regions, you paid a premium for an option you did not exercise.

Can I combine committed use discounts with sustained use discounts?

Yes. A committed use discount, resource-based or flexible, covers your committed baseline at the committed rate, and sustained use discounts apply automatically to eligible on-demand usage above the commitment, up to 30 percent for full-month usage. The common pattern is to commit the steady baseline and let sustained use discounts cover variable steady usage on top.

How does C3X help choose a committed use discount type?

C3X prices your Compute Engine footprint from Terraform before you deploy, so you can see the steady baseline that a commitment would cover. That helps you judge how stable your machine family and region footprint is and choose between a deeper resource-based commitment and a more flexible spend-based one on real numbers, sizing the commitment to genuine baseline usage rather than guesswork.

What to do next

Size your commitment to real baseline usage before you buy. C3X reads your Terraform and prices your resources against a live catalog. Start with the quickstart.

Try C3X on your own Terraform

Free and open source. No API key required. One command to install, one command to estimate.