A primary risk in multi-region designs is application unavailability if multiple regions fail simultaneously. When an application is deployed across two regions and both experience outages, the global load balancer remains available but has no healthy backends to distribute traffic to, which renders the service down. To mitigate this, critical applications must be deployed across more than two regions, and proper capacity planning must be performed to ensure either sufficient reserved capacity exists in all regions or the risks of emergency autoscaling are acceptable. Operational practices for disaster recovery testing, incident management, and post-incident verification are also essential to ensure rapid recovery.
When designing connectivity between regions and external networks, you must account for differences in how cloud providers handle routing. A Google Cloud VPC network supports global cross-region routing, while other cloud service providers might use regional VPCs with different Border Gateway Protocol scopes. The Cloud Router service automatically advertises routes with path preferences based on regional proximity, but this behavior depends on the VPC's configured dynamic routing mode. You may need to override these default preferences to achieve desired routing behavior, such as controlling which network path carries inter-regional traffic.
To manage latency and ensure flow symmetry where request and response traffic follow the same path, you should designate a preferred network for inter-regional communications using hot-potato and cold-potato routing strategies. Cold-potato routing in one domain paired with hot-potato routing in the peer domain keeps traffic on the originating network as long as possible, reducing latency and preventing issues with stateful security functions that asymmetric flows can cause.
For high availability within a single region, use active/active designs for intra-regional links. This design uses Equal Cost Multi-Path routing to aggregate the bandwidth of multiple paths, such as Cloud Interconnect connections, and use them simultaneously. Cloud Interconnect also supports LACP-aggregated links to achieve up to 200 Gbps of aggregate bandwidth per path. An active/passive design, where one link is a standby, is only necessary when on-premises network topologies combined with stateful security functions require it.
To increase availability to 99.99% for connections to on-premises networks, you must peer in multiple geographical regions. This involves connecting on-premises routers in different locations redundantly to the managed Cloud Router services in two different Google Cloud regions. This multi-region design provides backup connectivity in case one region fails.
You must configure external routers to control traffic paths and avoid site-to-site data transfer failures. Site-to-site data transfer via Network Connectivity Center is only supported in specific locations. If you maintain a connectivity resource in an unsupported location, you must influence BGP route selection to ensure traffic uses paths associated with Network Connectivity Center spokes.
To optimize for site-to-site data transfer, configure your external router to advertise prefixes with a lower Multi-Exit Discriminator value to the supported region and a higher MED to the unsupported region. This forces the Cloud Router to prefer and readvertise the route from the supported region, enabling site-to-site connectivity. To optimize for site-to-cloud traffic, advertise prefixes with the same MED to both regions, and the Cloud Router will choose the geographically closer path.
A Virtual Private Cloud network dynamic routing mode defines the operational scope for all Cloud Router instances deployed within that network. In regional dynamic routing mode, dynamic routes learned from external networks over Cloud Interconnect or Cloud VPN are only shared with subnets in the local Google Cloud region where the BGP session terminates. Conversely, global dynamic routing mode enables cross-region routing, allowing Cloud Routers in any region to advertise all VPC subnets to external networks and propagate learned on-premises prefixes to every subnet throughout the global VPC network.
Organizations implement regional dynamic routing when workloads require strict geographic boundaries or jurisdictional isolation where inter-regional traffic should not flow automatically. Global dynamic routing is selected when workloads span multiple regions and require simplified VPC-wide communication to external data centers without requiring manual route management.
Route propagation and path selection rely on Cloud Router advertising routes with predetermined path preferences calculated from regional proximity. When global dynamic routing is enabled, Cloud Router automatically assigns higher preference to local regional paths and lower preference to cross-region paths. Network engineers can override these default preferences using Multi-Exit Discriminator values or Cloud Router custom route advertisements to control ingress and egress path selection.
During Google Cloud route evaluation, the control plane processes route selection through a deterministic sequence. Policy-based routes and subnet routes are evaluated first, then the route selection model identifies the most specific destination prefix. Google Cloud evaluates the custom route type, preferring local dynamic and local static routes over Network Connectivity Center dynamic routes, and preferring NCC dynamic routes over peering custom routes. Unusable next hops and lower-priority routes are disregarded, leaving only routes with the highest priority for the final path selection.
Designing for high availability and disaster recovery requires aligning the VPC dynamic routing mode with physical connectivity topologies across one or more regions. A single-region deployment utilizes a managed Cloud Router distributed across multiple zones, paired with redundant Dedicated Interconnect or Partner Interconnect attachments in distinct edge availability domains to achieve a 99.9% service level agreement. These intra-regional connections commonly utilize Equal-Cost Multi-Path routing in an active/active configuration to aggregate link bandwidth or operate in an active/passive configuration when stateful security appliances require deterministic paths.
To achieve a 99.99% availability SLA, organizations must implement multi-region inter-domain routing using global dynamic routing paired with redundant connections across two distinct metropolitan areas and Google Cloud regions. If a regional outage or circuit failure occurs in the primary region, global dynamic routing automatically reroutes traffic through the secondary Cloud Router located in the alternate region, maintaining connectivity to application backends and supporting recovery time objectives without manual network reconfiguration.
Traffic engineering in multi-region hybrid networks requires managing path symmetry to prevent connection issues across stateful inspection firewalls and network virtual appliances. Flow symmetry is achieved by coordinating transit policies between Google Cloud and the peer external network, often pairing cold-potato routing in the Google Cloud domain with hot-potato routing in the external autonomous system. Under cold-potato routing, outbound traffic remains on the originating global Google network as long as possible until it reaches the exit interconnect closest to the destination.
Integrating dynamic routing with multi-region load balancing further optimizes traffic distribution and system resilience. Global external Application Load Balancers terminate user traffic at edge points of presence using Google Front Ends and route requests directly to healthy Compute Engine instances or managed instance groups across regions. When combined with global dynamic routing, internal tiers and database layers can fail over seamlessly across regions.
Multi-region inter-domain routing connects networks across multiple geographical areas to provide backup connectivity and increase availability. When you peer networks at multiple regions, the availability SLA can increase from 99.9% to 99.99% because traffic can fail over between regions when one region experiences an outage. Google Cloud VPC networks and Cloud Router both support global cross-region routing, which means routes learned in one region are automatically propagated to other regions.
Multi-region routing designs should address flow symmetry to ensure that outbound and return traffic follow the same network path. Flow asymmetry can cause problems with stateful security functions such as firewalls and intrusion detection systems. You can control the preferred network for inter-regional communications using hot-potato and cold-potato routing. Cold-potato routing keeps traffic on the originating network until it reaches the region closest to the destination, while hot-potato routing hands off traffic to the neighboring network as early as possible.
When designing inter-regional connectivity with Cloud Interconnect, you must configure connections as either active/active or active/passive. In an active/active design, traffic uses Equal Cost Multi-Path routing to aggregate the bandwidth of both paths and use them simultaneously for inter-domain traffic. This design can also use LACP-aggregated links to achieve up to 200 Gbps of aggregate bandwidth per path. In an active/passive design, one link remains on standby and only takes on traffic if the active link is interrupted.
For intra-regional links, an active/active design is recommended because it provides better bandwidth utilization and faster failover. However, certain on-premises networking topologies combined with stateful security functions may require an active/passive design to prevent traffic from being sent to a failed path that appears active. Cloud Router is instantiated across multiple zones within a region, which provides higher resiliency than a single router would provide.
Cloud Router is the service that manages dynamic routing between Google Cloud VPC networks and external networks such as on-premises data centers or other cloud service providers. Cloud Router automatically advertises routes with predetermined path preferences based on regional proximity, and this behavior depends on the configured dynamic routing mode of the VPC. You can choose between regional dynamic routing mode, where Cloud Router only learns and advertises routes within a single region, or global dynamic routing mode, where Cloud Router can learn and advertise routes across all regions.
When the default routing preferences do not match your requirements, you can override them using BGP attributes such as Multi-Exit Discriminator values. Cloud Router also supports Bidirectional Forwarding Detection for faster failure detection, which is essential for maintaining low recovery time objectives during regional outages.
Network Connectivity Center provides a hub-and-spoke model for managing connectivity between multiple VPC networks and external networks in a multi-region architecture. The hub serves as a centralized routing control plane that learns and distributes routes between VPC spokes and hybrid spokes. Using Network Connectivity Center, you can connect application VPCs, transit VPCs, and services-access VPCs in a mesh topology that enables any-spoke-to-any-spoke communication.
When designing multi-region architectures with Network Connectivity Center, you must consider site-to-site data transfer capabilities. This feature lets you connect external sites through the Google backbone network rather than requiring traffic to traverse the public internet. However, site-to-site data transfer is only supported in certain locations. If you need to maintain connectivity resources in unsupported regions, you can control traffic routing by configuring different MED values on your external router to prefer supported locations for site-to-site transfer while optimizing for latency in other scenarios.
HA VPN is the recommended method for implementing high-availability VPNs between regions and on-premises networks. When configured with BGP, HA VPN can achieve a 99.99% uptime SLA using either an active/active or active/passive tunnel configuration. In active/active mode, both tunnels carry traffic simultaneously and either can take over if the other fails. In active/passive mode, one tunnel remains idle until the active tunnel fails.
Cloud Interconnect provides dedicated physical connections between your on-premises network and Google Cloud, offering lower latency and higher bandwidth than VPN solutions. For multi-region disaster recovery, you can provision Dedicated Interconnect connections in multiple regions, each with redundant VLAN attachments in different edge availability domains within the same region. This design ensures that if one interconnection facility experiences an outage, traffic can fail over to the other facility within the same region, and if an entire region fails, traffic can fail over to the interconnected region.
In a hybrid multi-region environment, DNS resolution must work correctly across Google Cloud and external networks. Either Cloud DNS or an external on-premises DNS provider can handle lookups, with Cloud DNS being authoritative for Google Cloud zones and external DNS servers being authoritative for external zones. DNS forwarding must be enabled bidirectionally between Google Cloud and external networks, and firewalls must be configured to allow DNS resolution traffic.
For multi-region deployments using Shared VPC, cross-project binding of DNS zones enables segmentation and delegation of the DNS namespace to service project administrators. When external networks communicate with each other through Google Cloud in a transit scenario, the external DNS zones should forward requests directly to each other, with Google Cloud DNS only involved in forwarding DNS resolution traffic between external networks.
Disaster recovery connectivity strategies for multi-region architectures follow patterns that balance recovery time objectives and recovery point objectives. A hot disaster recovery pattern uses active/active deployments across regions with automatic failover, where global load balancers can route traffic to healthy regions when one region fails. A warm pattern uses standby resources that can be activated when needed, such as static sites in Cloud Storage or pre-provisioned Compute Engine instances. A cold pattern relies on recreating infrastructure from backups after a failure, which has the highest recovery time objective but lowest cost.
For database disaster recovery, Cloud SQL supports cross-region read replicas that can be promoted to primary instances during a regional outage. The advanced disaster recovery feature allows you to designate a disaster recovery replica that can be promoted immediately, with the original primary instance automatically becoming a replica of the new primary. A switchover operation can then restore the original topology with zero data loss by waiting for the replica to catch up before promoting it back to primary.
Regional dynamic routing restricts Cloud Router to learning and advertising routes only within the region where the BGP session terminates, while global dynamic routing propagates routes across all regions within the VPC network. Regional mode is used when workloads require strict geographic isolation, and global mode is used when workloads span multiple regions and need simplified VPC-wide communication.
Use active/active designs for intra-regional links when you need better bandwidth utilization and faster failover, as Equal Cost Multi-Path routing allows traffic to use multiple paths simultaneously. Use active/passive designs only when on-premises network topologies combined with stateful security functions require deterministic paths to prevent traffic from being sent to a failed path that appears active.
Flow symmetry ensures that outbound and return traffic follow the same network path, which is critical for stateful security functions like firewalls and intrusion detection systems. Asymmetric flows can cause these security devices to drop traffic. You achieve symmetry by coordinating transit policies between Google Cloud and the peer network, typically pairing cold-potato routing in Google Cloud with hot-potato routing in the external network.
Professional Cloud Network Engineer
Prepare and test your skills
Prepare and test your skills
StudyCerts
StudyCerts