Networking

Databricks provides a secure networking environment by default. You can configure additional networking features to control access to workspaces, secure connectivity between the control plane and compute planes, and protect connections to your data sources.

note

Databricks charges for networking costs when serverless workloads connect to customer resources, and when performance-intensive services egress data cross-region back to clients. See Understand Databricks networking costs.

Get started

Understand Databricks networking architecture and explore key concepts.

• • Network reference architectures
  • Learn about the control plane and compute plane architecture and choose a reference architecture for your network isolation requirements.
• • AWS PrivateLink
  • Establish private connections between your network and Databricks using AWS PrivateLink for enhanced security.
• • Understand data transfer and connectivity costs
  • Learn about data transfer pricing and optimize costs for network connectivity features.

Connectivity

Configure secure network connections for inbound access to workspaces and outbound connectivity from compute resources.

• • Front-end networking
  • Configure network access controls for users connecting to Databricks workspaces through the web interface and APIs.
• • Front-end PrivateLink
  • Enable private connectivity from your corporate network to Databricks workspaces using AWS PrivateLink.
• • Serverless compute plane networking
  • Configure secure network access between serverless compute resources and your data sources and services.
• • Private connectivity to AWS resources
  • Establish private connections from serverless compute to AWS services like S3, DynamoDB, and RDS.
• • Private connectivity to resources in your VPC
  • Connect serverless compute to resources running in your own VPC using private endpoints.
• • Manage private endpoint rules
  • Configure and manage private endpoint rules for serverless compute connectivity.
• • Classic compute plane networking
  • Learn about networking options for classic compute resources deployed in your virtual network.
• • Deploy Databricks in a customer-managed VPC
  • Host Databricks clusters in your own AWS VPC for enhanced network control.
• • VPC peering
  • Connect your Databricks VPC to other VPCs in your AWS account to access additional resources.
• • Back-end PrivateLink
  • Establish private connectivity between classic compute resources and the Databricks control plane.
• • Manage private access settings
  • Configure private connectivity settings at the account level for workspace deployment.
• • Manage VPC endpoint registrations
  • Register and manage VPC endpoints for private connectivity to Databricks services.
• • Configure Databricks S3 commit service-related settings
  • Optimize S3 write operations with network configurations for the Databricks S3 commit service.

Network security

Implement security controls to restrict and monitor network access.

• • What is serverless egress control?
  • Restrict outbound network connections from serverless compute resources to prevent data exfiltration and enforce compliance.
• • Manage network policies for serverless egress control
  • Create and manage network policies that define allowed egress connections from serverless compute.
• • IP access lists overview
  • Learn how to use IP access lists to control which IP addresses can access your Databricks workspaces.
• • IP access lists for workspaces
  • Configure workspace-level IP access controls to restrict access from approved networks.
• • IP access lists for the account console
  • Set account-level IP restrictions that apply across multiple workspaces for centralized security management.
• • Configure a firewall for serverless compute access
  • Use stable IP addresses to configure firewall rules for serverless compute connectivity.
• • Domain name firewall rules
  • Configure domain-based firewall rules to allow Databricks services through your network security controls.