Prepare and test your skills
Prepare and test your skills
Cloud VPN (opens in a new tab) is a Google Cloud Network Connectivity product that creates secure, encrypted tunnels over the public internet between a Virtual Private Cloud (VPC) network and an on-premises network or another cloud. Traffic flows from the on-premises network across the internet through the encrypted tunnel to the VPC network, and return traffic flows back through the same tunnel. For production work, the choice is between Cloud VPN and Cloud Interconnect, with Cloud Router (opens in a new tab) advertising routes between the environments.
High Availability (HA) VPN is the production-ready form of Cloud VPN that provides resilient, encrypted connectivity over the public internet without dedicated circuits. HA VPN carries a 99.99% service level agreement (SLA), which makes it suitable for production environments that cannot tolerate extended downtime. Compared to running an intermediary virtual machine (VM) as a VPN endpoint, HA VPN shifts uptime responsibility from the customer to Google infrastructure.
An intermediary VM creates a tradeoff between control and operational burden. The customer must maintain, secure, and keep the VM available, and a single VM failure can block all communication between environments. A Network Connectivity product removes that VM from the application path, so availability is managed by Google rather than by a customer-operated instance.
Cloud Router is the Google Cloud service that manages Border Gateway Protocol (BGP) routing between the VPC network and the on-premises router over Cloud VPN or Cloud Interconnect. Cloud Router advertises VPC routes to the on-premises environment and learns on-premises routes into the VPC network, so new subnets and topology changes propagate automatically without manual static routes. Cloud Router can also be configured to advertise custom IP ranges required by the private environment.
The VPN tunnels provide the encrypted transport while Cloud Router provides the control plane that decides where traffic flows. The tunnels connect the two network boundaries, and Cloud Router sessions running over those tunnels exchange reachability information in both directions. If the BGP session is not in the UP state, routes are not exchanged and traffic cannot be forwarded even if the tunnel is configured.
Exam tip: HA VPN over Cloud Interconnect requires its own Cloud Router for the HA VPN tier, separate from the Cloud Router for the Cloud Interconnect tier that exchanges VPN gateway prefixes.
Resilient hybrid architecture combines Cloud VPN with other Google Cloud services to keep applications serving traffic during partial or full environment failures. A hot high-availability pattern runs production workloads concurrently on-premises and in Google Cloud, with both environments serving production traffic and with database replication between the on-premises database server and the database server in Google Cloud. In this design, weighted Domain Name System (DNS) routing distributes client traffic across both environments, and failure in part of one environment causes traffic to be rerouted to the equivalent service in the other environment.
This pattern uses the following building blocks together:
When HA VPN is deployed over Cloud Interconnect, the design creates two operational tiers that each depend on a different router. The Cloud Interconnect tier includes the virtual local area network (VLAN) attachments and the Cloud Router for Cloud Interconnect, while the HA VPN tier includes the HA VPN gateways and tunnels and the Cloud Router for HA VPN.
Redundant hybrid topologies for high availability use managed connectivity and duplicated paths so no single circuit failure breaks the routed connection between an on-premises network and a VPC network. For production work, the recommended approach is a Network Connectivity product, either Cloud VPN (opens in a new tab) or Cloud Interconnect (opens in a new tab), with Cloud Router advertising routes between environments. This approach shifts uptime responsibility from a self-managed VM intermediary to Google and avoids the single point of failure of a single VM.
The production-level 99.99% availability configuration is the topology that qualifies for the 99.99% monthly uptime SLA for mission-critical applications with low tolerance for downtime. The 99.99% availability configuration requires four Cloud Interconnect connections across two Google Cloud regions. To verify the design, Google recommends using interconnect connection groups, which report whether the configuration meets the requirements for the intended availability level.
An on-premises network connects through four Cloud Interconnect connections split across two metros (Dallas and Ashburn) and separate edge availability domains into two Google Cloud regions (us-central1 and us-east4). Each region's Cloud Router terminates VLAN attachments and exchanges BGP routes with the VPC network, so if one metro, edge domain, or region fails, traffic continues over the remaining paths, qualifying for the 99.99% uptime SLA.
The production topology depends on redundant circuits in diverse locations:
Traffic flows from the on-premises router over the Interconnect connections to Cloud Router, and Cloud Router advertises VPC routes back to the on-premises network. If one metro, edge domain, or region is impaired, traffic can continue over the remaining paths to the other region.
Connection options for 99.99% availability are the three ways to meet the four-connection and two-region rules using Dedicated Interconnect, Partner Interconnect, or a mix of both. Dedicated Interconnect provides direct physical connections in a Google colocation facility, while Partner Interconnect provides connectivity through a supported service provider with a VLAN attachment in the customer project.
| Topology option | Composition | Placement rule |
|---|---|---|
| Four Dedicated connections | Four Dedicated Interconnect connections | Across two regions, different metros and separate edge availability domains |
| Four Partner attachments | Four Partner Interconnect VLAN attachments | Across two regions, different metros and separate edge availability domains |
| Mixed Dedicated and Partner | Two Dedicated connections in one region plus two Partner attachments in another region | Across two regions, different metros and separate edge availability domains |
Choose four Dedicated connections when direct colocation access is available, choose four Partner attachments when last-mile provider connectivity is needed, and choose the mixed option when the two regions have different access constraints.
Cloud Router is the dynamic routing service that exchanges routes between the VPC network and the on-premises router over Cloud VPN or Cloud Interconnect VLAN attachments. Cloud Router advertises VPC routes to the private environment, and the on-premises router advertises on-premises prefixes to the VPC network. When multiple paths exist to the same destination, forwarding follows the routing behavior managed through Cloud Router, so redundant circuits can serve as failover paths without manual route changes.
Single-circuit designs do not meet enterprise SLAs because they carry no uptime protection and can be impaired for long periods. A single Dedicated Interconnect connection, which is a Link Aggregation Control Protocol (LACP) bundle, carries No Uptime SLA, and a single Partner Interconnect VLAN attachment also carries No Uptime SLA. Single links or bundles can be drained for hours or days during planned or unplanned maintenance, and failures are handled at a support priority no higher than P3, meaning partially impaired service without a commitment to quick resolution or root-cause analysis.
The SLA also covers only the routed connection between the on-premises location and the VPC network, not the full application. For example, if Compute Engine instances are reachable through Cloud Interconnect, total service availability depends on the combined availability of both the Cloud Interconnect service and the Compute Engine service. Enterprise-grade availability therefore requires redundant circuits plus resilient compute, not interconnect redundancy alone.
Exam tip: A single Dedicated Interconnect connection and a single Partner Interconnect VLAN attachment both carry No Uptime SLA; 99.99% requires four connections across two regions with different metros and separate edge availability domains.
Dedicated Interconnect and Partner Interconnect are the two private options for Cloud Interconnect (opens in a new tab), which extends an on-premises network to one or more VPC networks without traversing the public internet. Dedicated Interconnect is a direct physical connection to Google, while Partner Interconnect reaches Google through a supported service provider. Both options provide access to all Google Cloud products and services from on-premises except Google Workspace, and both can reach supported APIs through Private Google Access from on-premises hosts.
Dedicated Interconnect (opens in a new tab) is a direct connection where traffic flows directly between the on-premises network and the VPC network, not through the public internet and not through a third-party provider. The organization installs and maintains its own routing equipment in a Google-supported colocation facility that serves the target Google Cloud regions, so the organization owns the physical meet with Google's edge network. Because the path is direct, Dedicated Interconnect is cost-effective when bandwidth demands are high and sustained.
Dedicated Interconnect scales through physical circuits and logical attachments. Connection capacity is delivered over one or more 10-Gbps, 100-Gbps, or 400-Gbps Ethernet circuits, with a maximum of 8 circuits per Dedicated Interconnect connection. On top of those circuits, flexible VLAN attachment capacities range from 50 Mbps to 50 Gbps. For routing, the organization configures BGP on both its on-premises routers and on Cloud Routers in Google Cloud, and Cloud Routers then apply learned on-premises routes as custom dynamic routes in the VPC network.
Partner Interconnect (opens in a new tab) is provider-managed private connectivity where traffic flows between the on-premises network and the VPC network through a supported service provider, still avoiding the public internet. The provider already connects directly to Google, so the organization does not need to install and maintain routing equipment in a colocation facility. Because providers offer more points of connectivity than the colocation facilities available for Dedicated Interconnect, Partner Interconnect is the option when the organization cannot physically meet Google's network.
Partner Interconnect is sized per VLAN attachment rather than per physical circuit. Each VLAN attachment offers flexible capacity from 50 Mbps to 50 Gbps, so the organization pays only for the capacity it needs. Adding more VLAN attachments or increasing the capacity of an existing attachment depends on the service provider's available capacity. Routing depends on the provider's service type: for Layer 2 connections the organization configures BGP on its on-premises routers and Cloud Routers, while for Layer 3 connections the Cloud Router configuration is fully automated and the on-premises router configuration follows the provider's documentation.
Choosing between Dedicated Interconnect and Partner Interconnect depends on bandwidth demand, physical proximity to Google, and willingness to manage colocation equipment. Choose Dedicated Interconnect for the most demanding data needs that justify 10-Gbps or larger circuits and direct control, and choose Partner Interconnect when a smaller 50 Mbps to 50 Gbps attachment is sufficient or when no colocation presence is possible. If traffic does not require a 10-Gbps, 100-Gbps, or 400-Gbps circuit, Partner Interconnect or Cloud VPN is the more practical fit.
| Option | Connectivity path | Bandwidth model | Setup and routing ownership |
|---|---|---|---|
| Dedicated Interconnect | Direct connection to Google; traffic flows directly between networks | 10-Gbps, 100-Gbps, or 400-Gbps circuits, up to 8 circuits per connection; VLAN attachments from 50 Mbps to 50 Gbps | Organization places routing equipment in a colocation facility; organization configures BGP on on-premises routers and Cloud Routers |
| Partner Interconnect | Connection through a supported service provider; traffic flows through the provider to Google | 50 Mbps to 50 Gbps per VLAN attachment; growth depends on provider capacity | Organization uses a supported provider with no colocation routers to maintain; BGP is customer-configured for Layer 2, Cloud Router is automated for Layer 3 |
Both options share the same encryption approach and differ in responsibility for availability and cost. Deploy HA VPN over Cloud Interconnect to encrypt traffic between the on-premises network and Google's network, which is supported for both Dedicated Interconnect and Partner Interconnect. For availability, Google offers an end-to-end SLA between the VPC network and on-premises network for Dedicated Interconnect using Google-defined topologies, while for Partner Interconnect Google covers the SLA between Google and the service provider and the provider may offer the end-to-end SLA. For cost, Partner Interconnect billing from Google is based on VLAN attachment capacity plus egress traffic from the VPC network to on-premises, with additional provider charges for carrying data across the provider network.
Exam tip: Choose Dedicated Interconnect when you can place equipment in a supported colocation facility and need 10-Gbps, 100-Gbps, or 400-Gbps circuits; choose Partner Interconnect when you cannot reach a colocation facility and need 50 Mbps to 50 Gbps through a supported provider.
Dedicated Interconnect is a direct physical connection to Google while Partner Interconnect connects through a supported service provider, and neither traverses the public internet. Choose Dedicated Interconnect when you can place routing equipment in a supported colocation facility and need 10-Gbps, 100-Gbps, or 400-Gbps circuits with up to 8 circuits per connection, and choose Partner Interconnect when you cannot reach a colocation facility and need 50 Mbps to 50 Gbps per Virtual Local Area Network (VLAN) attachment.
High Availability (HA) VPN is the production-ready form of Cloud VPN that provides resilient, encrypted connectivity over the public internet without dedicated circuits and carries a 99.99% service level agreement (SLA). It is suitable for production environments that cannot tolerate extended downtime because it shifts uptime responsibility from a customer-operated Virtual Machine (VM) intermediary to Google infrastructure and avoids a single Virtual Machine (VM) failure blocking all communication.
The 99.99% availability configuration requires four Cloud Interconnect connections across two Google Cloud regions with connection pairs in different metropolitan areas and separate edge availability domains. The requirement can be met with four Dedicated Interconnect connections, four Partner Interconnect Virtual Local Area Network (VLAN) attachments, or two Dedicated connections in one region plus two Partner attachments in another, with Cloud Router advertising routes between environments.
Cloud Router manages Border Gateway Protocol (BGP) routing between the Virtual Private Cloud (VPC) network and the on-premises router over Cloud VPN or Cloud Interconnect, advertising Virtual Private Cloud (VPC) routes outward and learning on-premises routes inward so topology changes propagate automatically. The tunnels provide the encrypted transport while Cloud Router provides the control plane that decides where traffic flows, and if the Border Gateway Protocol (BGP) session is not in UP state routes are not exchanged and traffic cannot be forwarded.
An enterprise is designing a secure hybrid networking solution to connect their on-premises data center to a Google Cloud Virtual Private Cloud (VPC) network over the public internet. The architecture must satisfy the following technical requirements:
How should the Cloud Architect implement this hybrid network connectivity?
Last updated