Red Hat OpenShift Container Platform

Red Hat OpenShift Container Platform

Our integrated solution for the Red Hat OpenShift Container Platform is a step forward to software modernization. The solution features implementations that run highly available and scalable micro-services, such as web servers, databases, java applications, etc.

What does this integrated solution offer?

  • OpenShift Container Platform offers accelerated application development and delivery with minimal configuration and management overhead supported by the latest hardware technologies like Intel® Xeon® Scalable Processors, NVMe storage, persistent memory and 25G ethernet.
  • Out of the box security from the operating system layer up to the container registries.
  • Additionally, the latest version of RedHat OpenShift has the option to boot the platform with FIPS (Federal Information Processing Standard) compliant encryption for handling extremely sensitive data and workloads.
  • Simple and highly automated installation and over-the-air updates utilized by Kubernetes operators.
  • The possibility of horizontal scaling by adding new racks to spread and increase workloads. Furthermore, the powerful hardware running the master nodes enables extending the cluster to around 1000 nodes without changing the control plane.
  • The latest NVMe storage is tested to comply with the needs of etcd, providing high endurance and low write latency.
  • Reliability and redundancy is achieved in both control level and data level.
  • OpenShift 4 is based on Ceph and Nooba storage which offers object, block and file capabilities. Combined with the high IOPS and microsecond grade latency offered by the latest NVMe storage units, it guarantees ultra-high performance.

Cluster specifications:

The cluster consists of 3 racks. One rack houses the master, infrastructure and storage nodes in addition to worker nodes. The other two racks house just worker and storage nodes.

  • 3 Masters nodes
  • 3 Infra nodes
  • 9 Storage nodes
  • 81 Worker nodes
  • Storage sled: OCP Accepted Lightning NVMe JBOF
  • Server model: OCP Accepted Tioga Pass
  • Networking: 2x25G, 1x1G per server
  • PDU: Max 18kW in 3+3
  • Memory: Pmem + DDR4
  • CPU: 2nd Generation Intel® Xeon® Scalable Processors
  • Storage type: Ceph on NVMe
  • Rack: OCP Inspired Open Rack v2

Master rack specifications:

  • 3 Masters nodes
  • 3 Infra nodes
  • 3 Storage nodes
  • 23 Worker nodes

Workers rack specifications:

  • 3 Storage nodes
  • 29 Worker nodes

Nodes specifications:

  • Master nodes have local storage for the OS and for the etcd cluster.
  • Storage nodes have local storage for the OS. Using Openshift Container Storage operator the storage nodes manage the Ceph storage, located in the lightening sled.
  • Worker nodes have local storage for the OS and use Ceph storage offered by the storage nodes as shared storage for the pods.
  • Worker nodes (all of part of them) can be supported with Pmem for (additional memory or as local storage) or a more powerful CPU for pods with intensive hardware usage.