A hybrid deployment operates a platform across a public cloud and one or more on-premises data centers. Anthos provides a common interface across these environments, where networking, resource provisioning, and operational responsibilities otherwise differ. section 10.1.2 section 10.1
A hybrid architecture typically combines:
- A Google Cloud deployment with one or more on-premises data centers. section 10.1.2
- Network connectivity through Cloud Interconnect, Cloud VPN, or the public internet. section 10.1.2
- A common platform interface for managing environments whose networks and resource-provisioning methods differ. section 10.1.2
- Load balancing and service-mesh routing that can send traffic to on-premises, Google Cloud, or other cloud services. section 7.3.2
Traffic Director can place Google Cloud load balancers in front of services wherever they run. Requests enter through a Google point of presence, pass through the load balancer and Envoy proxies, and reach services registered in the mesh. This enables cloud-native protection and delivery services, including Cloud Armor, Cloud CDN, Identity-Aware Proxy, and managed certificates. section 7.3.2
Migration without downtime
An application can split traffic between an on-premises service and its cloud replacement, gradually moving requests to the cloud while retaining a common entry point and service-mesh routing. section 7.3.2
Use the diagram to inspect how Dedicated Interconnect connects GCP with an on-premises data center. section 10.1.2
