You have a data processing pipeline that uses Cloud Dataproc to load data into BigQuery. A team of analysts works with the data using a Busing Intelligence (BI) tool running on Windows Virtual Machines (VMs) in Compute Engine. The BI tool is in use 24 hours a day, 7 days a week, and will be used increasingly over the coming years. The BI tool communicates to BigQuery only. Cloud Dataproc nodes are the main part of the GCP cost of this application. You want to reduce the cost without affecting the performance. What should you do?
A. Apply Committed Use Discounts to the BI tool VMs. Create the Cloud Dataproc cluster when loading data, and delete the cluster when no data is being loaded.
B. Apply Committed Use Discounts to the BI tool VMs and the Cloud Dataproc nodes. Create the Cloud Dataproc cluster when loading data, and delete the cluster when no data is being loaded.
C. Apply Committed Use Discounts to the BI tool VMs and the Cloud Dataproc nodes.
D. Apply Committed Use Discounts to the BI tool VMs.
Disclaimer
This is a practice question. There is no guarantee of coming this question in the certification exam.
Answer
A
Explanation
Committed Use Discount
– 1 to 3 years contract
– Give you a steep discount
– Ideal for workloads with predictable resource requirements
– It will bill you whether you use the resource or not
– Discounts differ based on the resources and time but could go as high as 50%-70%
A. Apply Committed Use Discounts to the BI tool VMs. Create the Cloud Dataproc cluster when loading data, and delete the cluster when no data is being loaded.
B. Apply Committed Use Discounts to the BI tool VMs and the Cloud Dataproc nodes. Create the Cloud Dataproc cluster when loading data, and delete the cluster when no data is being loaded.
(We will use Dataprocs occasionally, so applying CUD is not useful.)
C. Apply Committed Use Discounts to the BI tool VMs and the Cloud Dataproc nodes.
(We will use Dataprocs occasionally, so applying CUD is not useful.)
D. Apply Committed Use Discounts to the BI tool VMs.
(We are not optimizing the cost of Dataproc VMs which is the main part of the solution.)