Workload resource requirements

This section describes how to size an OpenShift-based container ecosystem cluster by using a sample cloud-native application. The following table shows a cloud-native inventory management application with a customized quotation generation system workload. Estimated memory, CPU core, I/O bandwidth, and storage requirements are indicative of resource requirements at times of peak load.

Table 5.              Estimated workload resource requirements by application type

The overall resource requirements are: 630 pods, 630 CPU cores, 2,047 GB RAM,
1.9 TB storage, and 130 Gbps aggregate network bandwidth.

Our calculations using the workload information from Table 5 take into account the following considerations:

• For each compute node configuration, it is recommended that you reserve four physical CPU cores per node for infrastructure I/O handling systems.
• Memory configuration is constrained to six DIMM modules per CPU socket (a total of 12 DIMM modules per node).
• DIMM module choices based on current trends for 2,933 MHz memory are
  16 GB, 32 GB, and 64 GB. The use of 16 GB DIMM modules results in a minimum node memory configuration of 192 GB.
• NIC options are: 2 x 25 GbE, 4 x 25 GbE, and 2 x 100 GbE.
• Overall compute node configuration options take account of the increased overall node workload handling capacity with processor and memory configuration. The configuration assumes that the compute nodes might be used over time for higher-performance workloads and that additional nodes will be installed to meet future growth in compute, storage, and network areas.

Compute node requirements example

Certain cluster design considerations apply to estimating the required number of compute nodes used. This section outlines these considerations.

• The number of pods required to be deployed is 630, which is clearly above the limit of 250 pods per node. The minimum number of nodes based on the limit of 250 pods per compute node is: 630 / 250 = 3 nodes.
• Table 5 provides estimates for the number of nodes that can be used to accommodate the projected workload. The cluster might require 40, 27, or 14 compute nodes, depending on the design of the node. Field experience recommends caution in the use of estimates for production use.

The following table shows the available configurations:

Table 6.              Calculated compute node alternate configurations based on Table 5 data

Controller node requirements

Our minimum recommended control-plane node configuration is a PowerEdge R640 server with dual Intel Gold 6226 CPUs and 192 GB RAM. As the Red Hat resource requirements show, this node is large enough for a 250-node cluster and higher. Dell Technologies recommends that you do not scale beyond 200 nodes, so the proposed reference design is adequate for nearly all deployments. The following table shows the sizing recommendations:

Table 7.              Control-plane node sizing guide

*Does not include provisioning of at least four cores per node for infrastructure I/O handling