Overview

To enable secure communication between Dataloop and your Google Cloud Platform (GCP) resources, two main integration methods are typically supported: Cross-Project Integration and Private Key Integration. These methods allow external platforms to authenticate and access GCP services such as Cloud Storage and BigQuery, depending on the required security level and use case.

1. Cross-Project Integration

Security Level: ⭐⭐⭐⭐⭐ (High)

Cross-Project Integration allows Dataloop to access GCP resources (like buckets or datasets) located in a different GCP project via IAM role delegation and resource sharing. This is a secure, scalable approach especially suited for enterprises with multiple GCP projects under the same organization.

🔑 Setup Includes:

• Creating a service account in the project that owns the data.
• Assigning it appropriate roles (permissions).
• Granting access to Dataloop's GCP service account by adding it as a member with specific roles (e.g., Storage Object Viewer, BigQuery Data Viewer) in your project or resource-level policy.

✅ Best For:

• Enterprise and multi-project environments
• Scenarios requiring resource segregation
• Centralized access management with minimal secret handling

2. Private Key Integration

Security Level: ⭐⭐☆☆☆ (Low)

Private Key Integration involves uploading a service account key file (JSON format) to Dataloop. This key enables Dataloop to authenticate as a service account and access specific GCP resources. It is fast to set up and often used in isolated or less complex environments.

🔑 Setup Includes:

• Creating a service account in your GCP project.
• Granting the service account the necessary roles (e.g., Storage Admin).
• Downloading the private key JSON file.
• Uploading that key securely to the Dataloop platform during integration.

⚠️ Considerations:

• Private keys must be stored securely and rotated periodically.
• Manual handling introduces a higher risk of credential exposure if not managed properly.

✅ Best For:

• Quick integrations
• Small or temporary projects
• Environments without organizational-level IAM policies