The following diagram is the reference architecture for vCloud Availability for vCloud Director 2.0 (disaster recovery to the cloud version of vCloud Availability), this architecture illustrates the required components between customer’s On-Premise datacenter and cloud provider’s vCloud Director based datacenter, also it shows network flows between all components of the vCloud Availability for vCloud Director solution. vCloud Availability solution is composed of infrastructure and solution components on both service provider and tenant sides.

small RA

(you can click here to see the original poster format of this reference architecture)

Although the poster format of reference architecture graph comes with the brief explanation for each component with small highlighted round icon (in orange color) ,  I’d like to add more description to those items here.

  1. vSphere Replication Appliance: vSphere Replication in appliance form factor provides the tenant side functionality. The on-premise vSphere Replication component combines both manager and replication server functionality. vSphere Replication Appliance can be extended with deployment of additional vSphere Replication Server instances for scale.
  2. vCTA: The on-premises vSphere Replication vCloud Tunneling Agent opens a control connection to the Cloud Proxy load balancer and then to a Cloud Proxy Cell. vCloud Tunneling Agent is the encryption module which communicates with the Cloud Proxy and is responsible for orchestrating tunnel creation for both to-the-cloud and from-the-cloud tunnels. It stops the to-cloud and from-cloud replication tunnels.
  3. ESXi: Replication data flows from the ESXi host on which the protected VM is running, to the vCloud Tunneling Agent which sends the data through the tunnel to the Cloud Proxy Service. The cloud proxy service redirects the replication data to the appropriate vRS server.
  4. Traffic to and from the cloud: Encryption of replication traffic is provided between vSphere Replication and Cloud Proxies endpoints.
  5. RabbitMQ: An open source message broker that implements AMQP. RabbitMQ is used to configure vCloud Director during the registration of the vCloud Replication Cloud Service. vCloud Availability for vCloud Director can utilize the same RabbitMQ cluster that is originally deployed with vCloud Director. The Cloud Service are routed through the messaging service. For high availability requirement, you can deploy two RabbitMQ load-balanced nodes running with RabbitMQ clustering enabled and mirrored queues are required.
  6. Cassandra: Cassandra NoSQL database cluster is used to store metadata about the replication, replicated VM instances, and infrastructure elements required to support the service.  vCloud Availability for vCloud Director can utilize the same Cassandra cluster that is deployed for the vCloud Director metrics database.
  7. Cloud Proxy: Cloud Proxy creates a public listening TCP port to which vCloud Tunneling Agents connect and communicate using secure web sockets. vCloud Director cell-like components are to tunnel replicated traffic to and from the cloud and establish control connections with on-premises vSphere Replication Appliance. Multiple instances can be deployed behind a load balancer to support scaling.
  8. vSphere Replication Cloud Service: It’s a tenant-aware replication manager that providers the required API for managing the service and all the components.
  9. vSphere Replication Management Server: It’s to oversee vSphere replication. The management server manages and monitors the replication process from tenant VMs to the service provider environment. The ratio of Replication Manager Server to the cloud resource VMware vCenter Server instances is one to one mapping. It is an existing component of Sphere Replication that has been updated for vCloud Availability.
  10. vSphere Replication Server: The Replication Service handles the replication process for each protected virtual machine. The replication server sends and receives the replication information and records the changes for each replicated VM. The ratio of Replication Server to the cloud resource vCenter Server instances is one to many.
  11. vCloud Availability Portal: It’s a virtual appliance that provides a graphical user interface for tenants to manage their replications when they have no access to their on-premises vSphere Web Client user interface. It provides a web user interface that is accessible on TCP port 8443 and requires a connection to the vCloud API. With a network load balancer supporting sticky sessions, the appliance can be deployed in highly available and horizontally scalable configuration.
  12. vCloud Availability Service Manager Portal: It’s a single virtual appliance that provides a graphical user interface for the service provider to monitor use and perform day 2 operations in the vCloud Availability environment. The portal offers a web-based user interface on TCP port 8443 and requires a vCloud API and RabbitMQ connection. The service provider can trigger storage migrations (relocation of replication to a different datastore), scrubbing stale migrations (for example, replications with missing placeholder vApps) and tenant impersonation feature (system administrator is redirected to tenant portal logged in as an organization user).

Hope this blog can help you understand more about the detail vCloud Availability for vCloud Director DR2C overall architecture, thanks!