Chapter 2. CIDR range definitions

To establish the network scope for cluster nodes in Red Hat OpenShift Service on AWS, specify an IP address range in the Machine Classless Inter-Domain Routing (CIDR) parameter. Defining this range ensures that all machines within the environment have valid, routable addresses for internal cluster communication.

This range must encompass all CIDR address ranges for your virtual private cloud (VPC) subnets. Subnets must be contiguous. A minimum range of 128 IP addresses, using the subnet prefix /25, is supported for single availability zone deployments. A minimum address range of 256 addresses, using the subnet prefix /24, is supported for deployments that use multiple availability zones.

The default is 10.0.0.0/16. This range must not conflict with any connected networks.

To allocate IP addresses for cluster services in Red Hat OpenShift Service on AWS, specify an IP address range in the Service Classless Inter-Domain Routing (CIDR) parameter. Defining this range ensures that internal services have a dedicated block of addresses for reliable communication without overlapping with node or pod networks.

Red Hat recommends, but this task is not mandatory, that the address block is the same between clusters. This does not create IP address conflicts.

The range must be large enough to accommodate your workload. The address block must not overlap with any external service accessed from within the cluster. The default is 172.30.0.0/16.

To allocate internal network addresses for cluster workloads in Red Hat OpenShift Service on AWS, specify an IP address range in the pod Classless Inter-Domain Routing (CIDR) field. Defining this range ensures that pods can communicate with each other reliably without overlapping with the node or service networks.

Red Hat recommends, but this task is not mandatory, that the address block is the same between clusters. This does not create IP address conflicts. The range must be large enough to accommodate your workload. The address block must not overlap with any external service accessed from within the cluster. The default is 10.128.0.0/14.

To allocate a dedicated pool of IP addresses for pods on each node in Red Hat OpenShift Service on AWS, specify the subnet prefix length in the hostPrefix parameter. Defining an appropriate prefix ensures that every machine has sufficient unique addresses to support its scheduled workloads without exhausting the cluster’s network resources.

For example, if you set the hostPrefix parameter to /23, each machine is assigned a /23 subnet from the pod CIDR address range. The default is /23, allowing 512 cluster nodes and 512 pods per node. Note that where 512 cluster nodes and 512 pods are beyond the maximum supported.

For example, if the host prefix is set to /23, each machine is assigned a /23 subnet from the pod CIDR address range. The default is /23, allowing 510 cluster nodes and 510 pod IP addresses per node.

Consider another example where you set the clusterNetwork.cidr parameter to 10.128.0.0/16, you define the complete address space for the cluster. This assigns a pool of 65,536 IP addresses to your cluster. If you then set the hostPrefix parameter to /23, you define a subnet slice to each node in the cluster, where the /23 slice becomes a subnet of the /16 subnet network. This assigns 512 IP addresses to each node, where 2 IP addresses get reserved for networking and broadcasting purposes. The following example calculation uses these IP address figures to determine the maximum number of nodes that you can create for your cluster:

65536 / 512 = 128

65536 / 512 = 128

You can use the Red Hat OpenShift Network Calculator to calculate the maximum number of nodes for your cluster.