When talking about moving telecom network functions and services to the cloud, the virtual network functions (VNFs) get most of the attention. The software part of the infrastructure—virtual infrastructure managers (VIMs), software-defined networking, and MANO platforms—get some attention, too. But the hardware?
The hardware is categorized as either compute, storage, or networking; labelled with abbreviations such as COTS (commercial, off-the-shelf systems), and dismissed as a commodity. This belief is reinforced by stories of how the cloud giants standardized on one design for their hardware and relentlessly optimized operations to exponentially increase their datacenter efficiency.
However, what is often overlooked is that although the hardware itself may have become a commodity, the infrastructure most certainly was not.
A common hardware pool, available for all workloads
You might be tempted to overlook the infrastructure in Ericsson's NFVi solution, too. As an open solution, it manages not only VNFs from multiple vendors, but it also uses hardware resources from multiple vendors. However, unlike other NFV infrastructures that rely on separate, statically-cabled configurations that are limited to hosting local workloads, we pre-cable our resources into a common hardware pool that is available to all workloads across the entire infrastructure.
That pool is open, and connects, through a common switching fabric, compute and storage resources from Ericsson and vendors such as Dell and HP. It then logically allocates those resources into virtual performance-optimized datacenters (vPODs), which can be rapidly created, expanded, reduced, or deleted without manual intervention.
This approach to a software-defined infrastructure provides real-world benefits under a variety of use cases. For instance, because the vPODs are logically isolated from each other, you can easily tune the global parameters of each VIM to match the requirements of the VNFs it hosts. As a result, you can deploy multiple VIM instances within the same infrastructure.
Only two of many use cases
If a VNF needs extra capacity, it can either add more resources from the common pool to the vPOD it is running on, or create a new vPOD. Either operation can be done in minutes. Hardware provisioning and infrastructure scaling has never been this fast or this easy.
These are only two use cases. There are many more, including simplified migration and updates, and unified hardware management across different organizations managing telecom VNFs, OSS/BSS, and IT. Together they extend the business agility and operational efficiencies of the cloud into the hardware layer, increasing the scalability, elasticity, and hardware utilization of your infrastructure.
Read about each one of them and the real-world benefits they provide in our latest white paper.