Software Defined Infrastructure upgraded with new Intel technology

Software Defined Infrastructure (SDI) extends the business agility and operational efficiency of the cloud into the hardware layer of NFVI (Network Functions Virtualization Infrastructure), which increases scalability, elasticity and utilization of datacenters. The Ericsson SDI system, the foundation of our market leading NFVI solution is now being upgraded with 2nd generation Intel® Xeon® Scalable processors, launched this week by Intel. Adding this new technology further strengthens Ericsson NFVI for central datacenter and edge computing deployments.

Tablet

Pre-integrated NFVI benefits 

Designing an NFVI solution involves several hardware and software components. To facilitate deployment and operations, Ericsson is promoting a pre-integrated solution which is tuned for performance and tested thoroughly. A key benefit of a pre-integrated solution is that when a new component is introduced, service providers know that it works as required and they understand the performance of the overall system. 

Pre-integration means that NFVI solutions can adopt new technology faster, compared to using a best-of-breed or decoupled approach. An example is enabling acceleration techniques throughout the cloud stack to increase the throughput of user-plane applications. Using a best-of-breed stack means that acceleration must be implemented in each layer of the stack. This approach isn’t a problem. But making this work across the solution between the various layers, including configuration, installation and operations is challenging. With a pre-integrated solution, acceleration is built into the full solution from the beginning. 

2nd generation Intel® Xeon® Scalable processors in Ericsson SDI 

The implementation of 2nd generation Intel® Xeon® Scalable processors in Ericsson SDI is being designed in the standard datacenter equipment Compute Sled Unit - CSU 02, and Compute Rack Unit - CRU 02. It will also be introduced in the short form factor CSU-S 02, which was recently demoed at Mobile World Congress 2019. CSU-S 02 can be deployed in central office environments and it can be co-located with radio equipment, making it ideal for applications that demand very low latency or early break-out to save bandwidth. Having the capability to use the same CPU type and version in both central and edge sites means an application can run in several locations without the need for parallel certification tracks, which makes it possible to move to the edge faster. This is a significant advantage. 

How does 2nd generation Intel® Xeon® Scalable processors make our solution better? Virtual network functions (VNFs) have different characteristics that result in different requirements on the infrastructure they’re running on. Since VNFs are sharing infrastructure resources in a cloud, it’s challenging to provide the optimal environment for all. EPC and 5G Core, including payload-centric applications, require a high-performance environment, while control-centric applications like IMS don’t have the same level of performance needs. 

A key feature of 2nd generation Intel® Xeon® Scalable processors is called Intel® Speed Select Technology. This feature makes it possible to allocate more frequency to the CPU cores handling the VNFs with high performance requirements, and less frequency to the ones that don’t need that much performance without deploying different servers for the different workloads. It’s a way of optimizing the total CPU capacity and of attaining a more homogenous, yet flexible set of deployed servers. This is an advantage, especially in edge computing deployments where small configurations must be able to handle several VNFs, often with different characteristics. 

Efficient hardware management 

Intel is also contributing to the Ericsson SDI system on the software level. To provide efficient hardware management, the two companies announced a collaboration in February about aligning and accelerating development efforts of the Ericsson SDI Manager software and Intel® Rack Scale Design. The intention is to bring new management functionality to the market and jointly broaden the ecosystem for Intel® Rack Scale Design compliant hardware.  

An additional outcome of the collaboration is to enable functionality in new hardware, both at central and edge sites and to make sure that the latter can operate autonomously - for example, at a link failure. This is an important capability to meet telco grade requirements on availability. Operational activities such as installation, upgrading, fault and performance management are also essential when deploying at the edge. These aspects are also addressed in the collaboration. 

Want to learn more about the Ericsson SDI system?

 

Download paper

Listen to Anders Vestergren, Head of Solution Line Software Defined Infrastructure explaining the role of the new 2nd generation Intel® Xeon® Scalable processor platform and how leveraging the performance and virtualization features built into the processors accelerate NFVI transformation in this podcast.

Read previous blog post on why pre-integrated NFVI makes a difference

Read more about what we do in NFV infrastructure


ABOUT THE CONTRIBUTOR
Henrik Bäckström
Henrik works as a Product Marketing Manager at Business Area Digital Services, focusing on cloud infrastructure products and network applications.
The Ericsson blog

In a world that is increasingly complex, we are on a quest for easy. At the Ericsson blog, we provide insight, news and opinion to help make complex ideas on technology, business and innovation simple. If you want to hear from us directly, please head over to our contact page.

Contact us