Cost-saving is already a topic for every enterprise moving to the cloud
As cloud development progresses, more and more enterprises experience the efficiency and convenience of migrating data to the cloud. "Moving to the cloud" is now a trend for many enterprises. However, how to spend the least money on the necessary cloud services is a challenge for every enterprise. No one wants to be taken advantage of; everyone wants their money to be well spent. Cloud providers have heard this wish and, from their perspective, they also want to accurately assess customer usage to achieve effective inventory management. By introducing cost-saving incentives, they encourage customers to purchase such plans. This not only saves costs for the clients but also allows cloud providers to effectively control inventory, creating a win-win situation. Today, I want to share with you the cloud cost-saving programs offered by GCP: CUD and SUD.

Committed Use Discounts (CUD) for GCP
CUD, short for Committed Use Discounts, is one of the cloud discount programs offered by GCP. Customers who commit to a certain level of usage over one or three years can save up to 57% on GCP resources compared to standard rates. CUD is more suitable for customers who primarily use virtual machines and Kubernetes or Compute Engine services, especially when the workload is stable and predictable. However, it's important to note that CUD offers discounts on specific resources under certain conditions. If the actual usage exceeds the committed level, GCP will charge the excess at standard rates.

CUD comes in two types—Spend-based and Resource-based. Spend-based CUD refers to a commitment by the customer to spend a certain amount of money on usage over one or three years, measured in dollars per hour but billed monthly. This scheme is applicable for using Cloud Run, Cloud SQL, Cloud Spanner, Memory Store, Compute Engine, VMware Engine, and Kubernetes Engine, with different products and commitment times offering different discounts. Resource-based CUD means that the customer commits to using a fixed amount of Compute Engine resources in a specific region over one or three years. This scheme is only applicable to Compute Engine products. Regardless of whether it's Spend-based or Resource-based CUD, customers can share discounts across all eligible projects under the same Cloud Billing account. Once CUD is purchased, prices will not be affected by any future product price changes by GCP until the commitment expires.

Sustained Use Discounts (SUD) for GCP
SUD, short for Sustained Use Discounts, is another cloud discount program offered by GCP. Customers who continuously use servers for a certain period can receive SUD. This is applicable for using general-purpose or predefined machine types, compute-optimized machine types, memory-optimized machine types, and the vCPU and memory of sole-tenant nodes, including the premium costs of single-tenant nodes, and all GPU devices (except for NVIDIA A100 and T4 GPU types). The unique aspect of SUD is that customers do not need to commit to anything; if they use a specific resource for more than 25% of a billing month, Compute Engine will automatically provide a discount for each incremental hour used. The longer the customer's workload runs, the more SUD discounts they can receive.

At the end of each month, Compute Engine automatically calculates the SUD obtained by the customer for that month, presenting the applicable SUD resources in a points format and directly offsetting costs. Customers can view the details in their Cloud Billing account. Moreover, SUD is divided into three types: SUD at different levels, where longer usage times earn higher discounts; Applying SUD, where Compute Engine analyzes and breaks down the number of vCPUs and memory used by all running virtual machines in the month to calculate the maximum discount; and SUD for GPUs, following the same concept, where Compute Engine analyzes the quantity of GPUs of the same model used by all running virtual machines in the month to calculate the maximum discount.

CUD & SUD Comparison
Finally, let's briefly compare the differences between CUD and SUD! The comparison is organized in the table below:

Source：

GCP 官方文檔 - 〈Committed use discounts〉

https://cloud.google.com/docs/cuds

GCP 官方文檔 - 〈Sustained use discounts for Compute Engine〉

https://cloud.google.com/compute/docs/sustained-use-discounts