Network Cloudification Requires a Village, Not a Vacuum.
By Dan Rodriguez, General Manager of Network Compute Division, Vice President of Intel Corporation
Since we are deeply invested in the success of our customers, the transformation of wireless, wireline and enterprise networks has been central to Intel’s evolution into a holistic data-centric company. Intel has increasingly met or exceeded the needs of our network customers as reflected in our growing share of the networking segment. When we
look at external reports alongside our own analysis, the consensus estimate is that investments in virtualized network servers will be roughly 10X greater than spending on fixed-function network appliances over the next few years. Against this backdrop we continue to modernize communication infrastructure to support 5G and edge computing through portfolio advancements, expanded ecosystem partnerships and deeper optimizations focused on key workloads.
Platform Innovation Built with the Voice of the Industry
One example of customer-focused innovation is the 2nd generation Intel® Xeon® Scalable processors with new memory hierarchy, connectivity and acceleration designed to excel in performance and infrastructure efficiency across a wide range of data-centric workloads. Intel also developed specialized network SKUs that allow the dynamic tuning and prioritization of the processor cores for maximum performance and power efficiency for virtual network functions. While performance is often critical for many workloads, the historical focus on service assurance and throughput in network operations makes being able to allocate performance for critical network functions an essential requirement for virtual networks.
Our deep ecosystem collaborations result in solutions designed with workload optimized features that deliver outstanding performance on thousands of real-world applications. The latest example is the Intel® FPGA Programmable Acceleration Card (Intel® FPGA PAC) N3000. This card allows customers to create custom tailored accelerated virtualized solutions for NFV, virtualized radio access network (vRAN) and edge workloads. We are enabling industry leading ecosystem partners to bring scalable and high performance NFV solutions to market at a faster pace than previously possible.
Embracing Network Edge, Exponential Growth and Innovation Opportunities
A robust ecosystem enables customers to quickly develop and port communications workloads to Intel architecture. This speed and flexibility will be fundamental as communications service providers (CommSPs), enterprises, cloud and IoT organizations rush to innovate across the network edge. Modernization of the radio access network (RAN), and specifically 5G wireless base stations, will be a linchpin of advanced 5G and edge services. The communications industry has already rewarded Intel’s technology and ecosystem investments with significant design wins in the RAN that will continue to support our customers’ needs in the future.
Another recent collaboration in network edge transformation is CenturyLink’s investment of several hundred million dollars in edge deployments, along with upgrades to its private cloud services on VMware Cloud Foundation (VCF) that include standardization on 2nd generation Intel Xeon Scalable processors. We are all watching new partnerships form between CommSPs and cloud service providers (CSPs) to bring the capabilities and scale of each organization to the enormous opportunities at the network and enterprise edge even as CommSPs accelerate adoption of cloud-native, programmable architectures.
Transforming 5G Core Network with Ecosystem Reference Architectures
While CommSPs are building the bridge to advanced 5G and edge services in stages, we are keenly focused on driving high-performance, virtualization of user-plane functions to support the bulk of 5G data traffic from the mobile core to the edge. In order to achieve a more fully disaggregated deployment model, we developed the Intel® Select Solutions for NFVI Forwarding Platform (FP) with our ecosystem partners to achieve the highest throughput possible on Intel® Xeon® Gold and Intel® Xeon® Platinum processors. The multi-node architecture supports Control and User Plane Separation (CUPS) strategy to distribute and scale user-plane functions across multiple systems to support sites with different requirements supporting multiple Terabits per second at the network core and the required speed and intelligence for the distributed edge services.
Intel has a long history of filling unmet industry needs through significant open source investments. The Intel Kubernetes Reference Implementation is delivered as part of the open source software initiative focused on cloudifiying the network. At SDN NFV World Congress we are delivering network services, performance efficiencies and automation in cloud native, edge and 5G architectures using Kubernetes Reference Implementation with F5 and Cloudify based on Intel’s 2nd generation
Xeon Scalable, Intel® QuickAssist technology, and Intel® Ethernet 700 NIC cards.
Network Cloudification Village on Display at The Hague
In my experience, customers buy systems inclusive of CPUs, memory, storage, software, networking, and accelerators. They care about overall solution value, the time to market and total cost of ownership (TCO) for their particular needs. Our customers are able to leverage server volume economics and access a rich and extensive tool chain to abstract network complexity and develop networking applications running on server-based architectures. When combined with the access to optimizations and pre-validated solutions from the broadest ecosystem in the world, the results we’ve seen are a testament to the value we’re delivering to the industry.
If you are attending the SDN NFV World Congress, we encourage you to join us and our partners at the Intel booth and various keynote presentations and panels. You can get real-time updates through my Twitter handle, @DanRod2000,@IntelNetwork and intel.com/network.