Grid Network topology

The simplest network topology is created by configuring the Grid Network only.

When you configure the Grid Network, you establish the host IP address, subnet mask, and Gateway IP address for the eth0 interface for each grid node.

During configuration, you create a list of subnets called the Grid Network Subnet List (GNSL). This list contains the list of subnets in the Grid Network and the external subnets interfacing to the Grid Network. At installation, the Grid Network interface auto-generates static routing tables for all subnets in the GNSL and adds a default route if the Grid Network has a gateway and there is no Client Network.

Note: Optionally, you can configure high availability (HA) groups of Admin and Gateway Nodes for the Grid Network. If HA groups are configured, the virtual IP addresses (VIPs) of the HA groups must be within the same subnet as the interface they are being attached to and must follow the same routing rules.

In this example, all grid traffic shares the same network, including traffic related to S3 and Swift client requests, object management operations, and administrative functions.

Note: This topology is appropriate for single site deployments that are not externally available, and proof-of-concept or test deployments. When possible, the Grid Network should be used exclusively for internal traffic. Both the Admin Network and the Client Network have additional firewall restrictions that block external traffic to internal services. Using the Grid Network for external client traffic is supported, but this use offers fewer layers of protection.