Valid Professional-Cloud-Architect Dumps shared by ExamDiscuss.com for Helping Passing Professional-Cloud-Architect Exam! ExamDiscuss.com now offer the newest Professional-Cloud-Architect exam dumps, the ExamDiscuss.com Professional-Cloud-Architect exam questions have been updated and answers have been corrected get the newest ExamDiscuss.com Professional-Cloud-Architect dumps with Test Engine here:
Your team plans to use Vertex AI to develop and deploy machine learning models for various use cases for fraud detection, product recommendations, and customer churn prediction. You want to enhance the security posture of the Vertex AI and Workbench environment by restricting data exfiltration. What should you do?
Correct Answer: C
According to the Google Cloud Security Foundations Guide and VPC Service Controls (VPC-SC) documentation, the most effective way to prevent data exfiltration is to establish a service perimeter . This perimeter creates a virtual boundary that prevents sensitive data from being moved to unauthorized projects or outside the Google Cloud network. For the Vertex AI environment specifically, the two critical services that must be protected are aiplatform. googleapis.com (the core Vertex AI API) and notebooks.googleapis.com (which governs Vertex AI Workbench instances). Including these in a service perimeter ensures that data within the Cymbal Retail project cannot be copied to external Cloud Storage buckets or other unauthorized APIs, even if a user has valid IAM permissions. Option A is incorrect because ml.googleapis.com is the legacy AI Platform API, and document.googleapis. com (Document AI) does not cover the broader Vertex AI or Workbench environment. Option B (VPC Flow Logs) provides visibility but is a detective control, not a preventative one against exfiltration. Option D (Private Google Access) allows internal IPs to reach Google APIs but does not restrict where that data can go once the connection is made. Therefore, Option C provides the robust, enterprise-grade security required for sensitive data like fraud detection and customer churn models.