A typical telecommunications (telecom) company sells telecom-oriented applications as a service to its consumers. Telecom use-case requirements vary depending on the virtual network functions (VNFs) that are being serviced. These functions include:
- Content delivery network (CDN)
- Edge infrastructure and towers of power
- NFV management and operations (NFV-MANO)
- Software-defined networking (SDN) and SD-WAN management
- Radio access networks (RAN) and 5G, and their component service infrastructures
- Multiaccess Edge Computing (MEC)
- Core network and 5G Next Generation Core (NGC)
This use case identifies some key design factors for a telco container platform.
Online video consumption has grown in recent years. High-quality video delivery over public networks requires a CDN. To handle growth, many operators are considering the virtualization of the CDN, giving them the ability to scale CDN on demand. CDN virtualization permits simple provisioning and sharing of resources with other telecom services, simplifying operations and avoiding costly dedicated infrastructure.
The following figure shows a virtual CDN (vCDN):
Figure 13. Virtual content delivery network (vCDN)
A vCDN stack requires the following principal capabilities:
- Large application storage space to store video and other files
- High-speed and low-latency network options to serve the content
- Rapid ramp-up of on-demand processing capacity
OpenShift Container Platform 4.6 on PowerEdge hardware platforms meets these demands by providing the following capabilities:
- CSI storage drivers for Dell Unity, PowerMax, PowerFlexOS, and PowerScale (formerly Isilon) storage systems are available and have been certified by Red Hat. These drivers can be integrated into your OpenShift Container Platform deployment using the new CSI plug-ins.
- High-speed (25 GbE/100 GbE) network interfaces of Dell Technologies server and switch portfolios meet the networking needs of network I/O-intensive applications.
- OpenShift Data Foundation (based on Ceph) is supported as part of OpenShift Container Platform 4.6.
- Multus CNI plug-in support by which additional networks can be added to each container so that the container can meet capacity needs on targeted networks.
- SR-IOV is now natively supported by OpenShift Container Platform. Red Hat provides an SR-IOV Operator over OperatorHub, enabling administrators to manage virtual functions on nodes through Kubernetes CRDs.
- DPDK is available in OpenShift Container Platform 4.6 as a Technical Preview feature. Currently, DPDK is not supported by Red Hat in production environments.
- OpenShift Container Platform now supports running workloads inside VMs that are natively managed and deployed by OpenShift Container Platform.
- For low-latency workloads, you can configure compute nodes running Red Hat CoreOS to use the real-time kernel, which provides consistent low-latency performance.
- Compute nodes can also be deployed at the edge, managed by a central cluster located in a data center.
- Small footprint clusters (three-node clusters) can also be deployed at the edge for production environments.
- Local storage on edge nodes can be provisioned for use by applications, allowing edge nodes to operate independently of an external storage array, at the cost of resiliency.
- Telco applications generally use huge pages. In OpenShift Container Platform, applications can allocate and consume huge pages.
- OpenShift Container Platform 4.6 includes support for IPv6.
Container ecosystem clusters for large deployments can span across multiple racks. We highly recommend using leaf-spine networking when scaling to more than three racks per cluster. The deployment of large clusters requires significant modification of the Ansible playbooks that were generated to facilitate large-scale deployment.