In order to provision cloud clusters, you must first deploy a network into your cloud account. A network is a networking stack deployed into your cloud service provider (CSP) that clusters run on top of. Networks are managed per cloud account, and you can deploy multiple networks within a single account.
Below, you can find specific details about the implementation of networks for each CSP.
Deploying a network provisions the following resources on AWS:
The schematic below shows an ACTIVATE-deployed cluster in AWS Cloud.
The controller and compute nodes are shown in Availability Zone us-east-1a inside the us-east-1 Region. When using the default networking mode, compute nodes in a region use a single NAT gateway for outbound internet connectivity. This NAT gateway is pre-provisioned as part of the network. The AMI used for the controller and compute nodes is inside the Parallel Works AWS account.
Expense Information
Depending on the networking mode, a NAT gateway will be created resulting in an additional monthly charge in your AWS account. For more information, please see this AWS pricing list.
A diagram of ACTIVATE clusters in AWS Cloud
Deploying a network on Azure provisions a similar set of networking resources, including a virtual network, subnets, NAT gateway, and associated security rules.
A diagram of ACTIVATE clusters in Azure Cloud
Deploying a network on Google Cloud provisions a VPC network, subnets, firewall rules, and Cloud DNS resources in the specified project and region.
For OpenStack environments, network configuration is specific to your OpenStack deployment. Networks are created using your provider's available networking resources, regions, and flavors.
