Committed Use Discounts

Committed use discounts, usually called CUDs, are Google Cloud's equivalent of reserved instances and savings plans. A customer commits to use a certain amount of resources, such as virtual CPUs and memory in a region, or to spend a certain amount per hour on eligible services, for a one-year or three-year term. In return, Google Cloud charges lower rates for the committed usage. Like other commitment discounts, CUDs apply automatically to matching usage, and the customer pays for the commitment whether or not it is fully used.

Google Cloud offers two main kinds: resource-based commitments, which commit to specific quantities of compute resources, and spend-based commitments, which commit to an hourly spend on particular services. For IT financial management, committed use discounts are a key rate optimization tool for organizations using Google Cloud, and they need the same careful sizing, monitoring and allocation as commitments on other platforms.

Definition of Committed Use Discounts

Committed use discounts are contracts with Google Cloud in which a customer agrees to a minimum level of resource usage or spending over a one-year or three-year term, in exchange for discounted pricing on that usage. Three-year terms generally offer larger discounts than one-year terms.

CUDs are distinct from sustained use discounts, which Google Cloud applies automatically to certain compute resources that run for a significant part of the month, without any commitment.

Resource-Based Commitments

Resource-based commitments apply to Compute Engine resources, such as virtual CPUs, memory, GPUs and local SSDs, in a specific region and machine series. The customer commits to a quantity, for example 400 vCPUs and 1,600 GB of memory, and pays the discounted rate for that quantity throughout the term.

The discount applies to any matching usage in the region and project scope, regardless of which individual virtual machines are running. Organizations can share commitments across projects within a billing account.

Because they are tied to a machine series and region, resource-based commitments work best for stable workloads whose architecture is unlikely to change during the term.

Spend-Based Commitments

Spend-based commitments commit to a minimum hourly spend on eligible services, such as Cloud SQL, Google Kubernetes Engine or flexible Compute Engine usage. The discount applies to usage up to the committed amount.

These commitments are more flexible, since they apply across machine types and, in some cases, regions, similar to savings plans on other platforms.

They usually offer smaller discounts than resource-based commitments in exchange for that flexibility.

Importance of Committed Use Discounts in IT Financial Management

For organizations running significant workloads on Google Cloud, steady compute and database usage paid at on-demand rates is a major source of avoidable cost. CUDs reduce that cost.

They also introduce committed spending that must be managed. Once purchased, a commitment must be paid for the full term, so decisions about how much to commit have consequences that last for years.

Getting the balance right between savings and flexibility is as important on Google Cloud as on any other platform.

Savings and Coverage

CUDs can reduce the cost of steady usage substantially compared with on-demand rates. Combined with sustained use discounts and other pricing options, they form the core of rate optimization on Google Cloud.

As with reserved instances, the savings depend on how much eligible usage the commitments cover, measured by commitment coverage, and how fully they are used.

Organizations using several cloud providers should track coverage and savings on each, so that commitment strategies are consistent. Differences in how each provider structures its commitments make standardized data, such as the FOCUS specification, especially useful here.

Risk of Unused Commitments

Resource-based commitments are tied to specific machine series and regions. If workloads move to a different series, region or service during the term, the commitment may go unused, lowering commitment utilization.

Spend-based commitments reduce this risk but still require the committed spend to be reached each hour.

Committed amounts should be treated as fixed costs in budgets and forecasts for the length of the term.